Received: February 15, 2022. Revised: May 16, 2022. Accepted: June 1, 2022 © The Author(s) 2022. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Human Molecular Genetics, 2022, Vol. 00, 00, 1–18 https://doi.org/10.1093/hmg/ddac132 Advance access publication date: 16 June 2022 Original Article Novel genes and sex differences in COVID-19 severity Raquel Cruz 1,2,3,4,†, Silvia Diz-de Almeida4,†, Miguel López de Heredia2, Inés Quintela1, Francisco C. Ceballos5, Guillermo Pita6, José M. Lorenzo-Salazar7, Rafaela González-Montelongo7, Manuela Gago-Domínguez8,3, Marta Sevilla Porras2,9, Jair Antonio Tenorio Castaño2,9,10, Julian Nevado2,9,10, Jose María Aguado11,12,13,14, Carlos Aguilar15, Sergio Aguilera-Albesa16,17, Virginia Almadana18, Berta Almoguera19,2, Nuria Alvarez6, Álvaro Andreu-Bernabeu20,13, Eunate Arana-Arri21,22, Celso Arango20,23,13, María J. Arranz24, Maria-Jesus Artiga25, Raúl C. Baptista-Rosas26,27,28, María Barreda-Sánchez29,30, Moncef Belhassen-Garcia31,32, Joao F. Bezerra33, Marcos A.C. Bezerra34, Lucía Boix-Palop35, María Brion 36,37, Ramón Brugada38,39,37,40, Matilde Bustos41, Enrique J. Calderón 42,43,44, Cristina Carbonell45,32, Luis Castano21,46,2,47,48, Jose E. Castelao49, Rosa Conde-Vicente50, M. Lourdes Cordero-Lorenzana51, Jose L. Cortes-Sanchez52,53, Marta Corton19,2, M. Teresa Darnaude54, Alba De Martino-Rodríguez55,56, Victor del Campo-Pérez57, Aranzazu Diaz de Bustamante54, Elena Domínguez-Garrido58, Andre D. Luchessi59, Rocío Eiros60, Gladys Mercedes Estigarribia Sanabria61, María Carmen Fariñas62,63,64, Uxía Fernández-Robelo65, Amanda Fernández-Rodríguez5,14, Tania Fernández-Villa66, Belén Gil-Fournier67, Javier Gómez-Arrue55,56, Beatriz González Álvarez55,56, Fernan Gonzalez Bernaldo de Quirós68, Javier González-Peñas20,13,23, Juan F. Gutiérrez-Bautista 69, María José Herrero70,71, Antonio Herrero-Gonzalez72, María A. Jimenez-Sousa5,14, María Claudia Lattig73,74, Anabel Liger Borja75, Rosario Lopez-Rodriguez19,2, Esther Mancebo76,77, Caridad Martín-López75, Vicente Martín78,43, Oscar Martinez-Nieto79,74, Iciar Martinez-Lopez80,81, Michel F. Martinez-Resendez52, Angel Martinez-Perez82, Juliana F. Mazzeu83,84,85, Eleuterio Merayo Macías86, Pablo Minguez19,2, Victor Moreno Cuerda87,88, Vivian N. Silbiger59, Silviene F. Oliveira83,89,90,91,92, Eva Ortega-Paino25, Mara Parellada20,23,13, Estela Paz-Artal76,77,93, Ney P.C. Santos94, Patricia Pérez-Matute95, Patricia Perez96, M. Elena Pérez-Tomás29, Teresa Perucho97, Mel Lina Pinsach-Abuin38,37, Ericka N. Pompa-Mera98, Gloria L. Porras-Hurtado99, Aurora Pujol100,2,101, Soraya Ramiro León67, Salvador Resino5,14, Marianne R. Fernandes94,102, Emilio Rodríguez-Ruiz103,3, Fernando Rodriguez-Artalejo104,105,43,106, José A. Rodriguez-Garcia107, Francisco Ruiz Cabello69,108,109, Javier Ruiz-Hornillos110,111,112, Pablo Ryan113,114,115, José Manuel Soria82, Juan Carlos Souto116, Eduardo Tamayo117,118, Alvaro Tamayo-Velasco119, Juan Carlos Taracido-Fernandez72, Alejandro Teper120, Lilian Torres-Tobar121, Miguel Urioste122, Juan Valencia-Ramos123, Zuleima Yáñez124, Ruth Zarate125, Tomoko Nakanishi 126,127,128,129,130, Sara Pigazzini131,132, Frauke Degenhardt 133,134, Guillaume Butler-Laporte 135,128, Douglas Maya-Miles136,137, Luis Bujanda138,137, Youssef Bouysran139, Adriana Palom140,141,142, David Ellinghaus143,133, Manuel Martínez-Bueno144, Selina Rolker145, Sara Amitrano146, Luisa Roade137,140,147, Francesca Fava148,149,150, Christoph D. Spinner 151, Daniele Prati152, David Bernardo153,137, Federico Garcia154,155, Gilles Darcis156,157, Israel Fernández-Cadenas158, Jan Cato Holter159,160, Jesus M. Banales161,137, Robert Frithiof162, Stefano Duga163,164, Rosanna Asselta163,164, Alexandre C. Pereira165, Manuel Romero-Gómez136,137, Beatriz Nafría-Jiménez166, Johannes R. Hov167,160,168, Isabelle Migeotte169,139, Alessandra Renieri 148,149,150, Anna M. Planas170,171, Kerstin U. Ludwig 145, Maria Buti137,140,147, Souad Rahmouni156, Marta E. Alarcón-Riquelme144,172, Eva C. Schulte173,174,175, Andre Franke133,134, Tom H. Karlsen167,160,176, Luca Valenti177,178, Hugo Zeberg179,180, Brent Richards181,128,182, Andrea Ganna132,183, Mercè Boada184,185, Itziar de Rojas184,185, Agustín Ruiz184,185, Pascual Sánchez-Juan186, Luis Miguel Real187, SCOURGE Cohort Group*, HOSTAGE Cohort Group*, GRA@CE Cohort Group*, Encarna Guillen-Navarro29,188,189,190, Carmen Ayuso19,2, Anna González-Neira6, José A. Riancho62,63,64, Augusto Rojas-Martinez 191, Carlos Flores7,192,193,194,‡, Pablo Lapunzina2,9,10,‡ and Angel Carracedo1,2,3,4,8,‡ 1Centro Nacional de Genotipado (CEGEN), Universidade de Santiago de Compostela, 15706 Santiago de Compostela, Spain 2Centre for Biomedical Network Research on Rare Diseases (CIBERER), Instituto de Salud Carlos III, 28029 Madrid, Spain 3Instituto de Investigación Sanitaria de Santiago (IDIS), 15706 Santiago de Compostela, Spain D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 2 | Human Molecular Genetics, 2022, Vol. 00, No. 00 4Centro Singular de Investigación en Medicina Molecular y Enfermedades Crónicas (CIMUS), Universidade de Santiago de Compostela, 15782 Santiago de Compostela, Spain 5Unidad de Infección Viral e Inmunidad, Centro Nacional de Microbiología (CNM), Instituto de Salud Carlos III (ISCIII), 28220 Madrid, Spain 6Spanish National Cancer Research Centre, Human Genotyping-CEGEN Unit, 28029 Madrid, Spain 7Genomics Division, Instituto Tecnológico y de Energías Renovables, 38600 Santa Cruz de Tenerife, Spain 8Fundación Pública Galega de Medicina Xenómica, Sistema Galego de Saúde (SERGAS), 15706 Santiago de Compostela, Spain 9Instituto de Genética Médica y Molecular (INGEMM), Hospital Universitario La Paz-IDIPAZ, 28046 Madrid, Spain 10ERN-ITHACA-European Reference Network 11Unit of Infectious Diseases, Hospital Universitario 12 de Octubre, Instituto de Investigación Sanitaria Hospital 12 de Octubre (imas12), 28041 Madrid, Spain 12Spanish Network for Research in Infectious Diseases (REIPI RD16/0016/0002), Instituto de Salud Carlos III, 28029 Madrid, Spain 13School of Medicine, Universidad Complutense, 28040 Madrid, Spain 14Centro de Investigación Biomédica en Red de Enfermedades Infecciosas (CIBERINFEC), Instituto de Salud Carlos III, 28029 Madrid, Spain 15Hospital General Santa Bárbara de Soria, 42005 Soria, Spain 16Pediatric Neurology Unit, Department of Pediatrics, Navarra Health Service Hospital, 31008 Pamplona, Spain 17Navarra Health Service, NavarraBioMed Research Group, 31008 Pamplona, Spain 18Hospital Universitario Virgen Macarena, Neumología, 41009 Seville, Spain 19Department of Genetics & Genomics, Instituto de Investigación Sanitaria-Fundación Jiménez Díaz University Hospital - Universidad Autónoma de Madrid (IIS-FJD, UAM), 28040 Madrid, Spain 20Department of Child and Adolescent Psychiatry, Institute of Psychiatry and Mental Health, Hospital General Universitario Gregorio Marañón (IiSGM), 28007 Madrid, Spain 21Biocruces Bizkai HRI, 48903 Barakaldo, Bizkaia, Spain 22Cruces University Hospital, Osakidetza, 48903 Barakaldo, Bizkaia, Spain 23Centre for Biomedical Network Research on Mental Health (CIBERSAM), Instituto de Salud Carlos III, 28029 Madrid, Spain 24Fundació Docència I Recerca Mutua Terrassa, 08221 Terrassa, Spain 25Spanish National Cancer Research Center, CNIO Biobank, 28029 Madrid, Spain 26Hospital General de Occidente, 45170 Zapopan, Jalisco, Mexico 27Centro Universitario de Tonalá, Universidad de Guadalajara, 45425 Tonalá, Jalisco, Mexico 28Centro de Investigación Multidisciplinario en Salud, Universidad de Guadalajara, 45425 Tonalá, Jalisco, Mexico 29Instituto Murciano de Investigación Biosanitaria (IMIB-Arrixaca), 30120 Murcia, Spain 30Universidad Católica San Antonio de Murcia (UCAM), 30120 Murcia, Spain 31Hospital Universitario de Salamanca-IBSAL, Servicio de Medicina Interna-Unidad de Enfermedades Infecciosas, 37007 Salamanca, Spain 32Universidad de Salamanca, 37007 Salamanca, Spain 33Escola Tecnica de Saúde, Laboratorio de Vigilancia Molecular Aplicada, 68515-000 Pará, Brazil 34Federal University of Pernambuco, Genetics Postgraduate Program, Recife 50670-907, PE, Brazil 35Hospital Universitario Mutua Terrassa, 08221 Terrassa, Spain 36Instituto de Investigación Sanitaria de Santiago (IDIS), Xenética Cardiovascular, 15706 Santiago de Compostela, Spain 37Centre for Biomedical Network Research on Cardiovascular Diseases (CIBERCV), Instituto de Salud Carlos III, 28029 Madrid, Spain 38Cardiovascular Genetics Center, Institut d’Investigació Biomèdica Girona (IDIBGI), 17190 Girona, Spain 39Medical Science Department, School of Medicine, University of Girona, 17190 Girona, Spain 40Hospital Josep Trueta, Cardiology Service, 17190 Girona, Spain 41Institute of Biomedicine of Seville (IBiS), Consejo Superior de Investigaciones Científicas (CSIC)- University of Seville- Virgen del Rocio University Hospital, 41013 Seville, Spain 42Departemento de Medicina, Hospital Universitario Virgen del Rocío, Universidad de Sevilla, 41013 Seville, Spain 43Centre for Biomedical Network Research on Epidemiology and Public Health (CIBERESP), Instituto de Salud Carlos III, 28029 Madrid, Spain 44Instituto de Biomedicina de Sevilla, 41013 Seville, Spain 45Hospital Universitario de Salamanca-IBSAL, Servicio de Medicina Interna, 37007 Salamanca, Spain 46Osakidetza, Cruces University Hospital, 48903 Barakaldo, Bizkaia, Spain 47Centre for Biomedical Network Research on Diabetes and Metabolic Associated Diseases (CIBERDEM), Instituto de Salud Carlos III, 28029 Madrid, Spain 48University of Pais Vasco, UPV/EHU, 48903 Bizkaia, Spain 49Oncology and Genetics Unit, Instituto de Investigacion Sanitaria Galicia Sur, Xerencia de Xestion Integrada de Vigo-Servizo Galego de Saúde, 36312 Vigo, Spain 50Hospital Universitario Río Hortega, 47012 Valladolid, Spain 51Servicio de Medicina intensiva, Complejo Hospitalario Universitario de A Coruña (CHUAC), Sistema Galego de Saúde (SERGAS), 15009 A Coruña, Spain 52Tecnológico de Monterrey, 64718 Monterrey, Mexico 53Otto von Guericke University, Departament of Microgravity and Translational Regenerative Medicine, 39106 Magdeburg, Germany 54Hospital Universitario Mostoles, Unidad de Genética, 28935 Madrid, Spain 55Instituto Aragonés de Ciencias de la Salud (IACS), 50009 Zaragoza, Spain 56Instituto Investigación Sanitaria Aragón (IIS-Aragon), 50009 Zaragoza, Spain 57Preventive Medicine Department, Instituto de Investigacion Sanitaria Galicia Sur, Xerencia de Xestion Integrada de Vigo-Servizo Galego de Saúde, 36312 Vigo, Spain 58Unidad Diagnóstico Molecular. Fundación Rioja Salud, 26006 La Rioja, Spain 59Universidade Federal do Rio Grande do Norte, Departamento de Analises Clinicas e Toxicologicas, 59012-570 Natal, Brazil 60Hospital Universitario de Salamanca-IBSAL, Servicio de Cardiología, 37007 Salamanca, Spain 61Instituto Regional de Investigación en Salud-Universidad Nacional de Caaguazú, HH36+J3Q Caaguazú, Paraguay 62IDIVAL, 39008 Cantabria, Spain 63Universidad de Cantabria, 39008 Cantabria, Spain 64Hospital U M Valdecilla, 39008 Cantabria, Spain 65Urgencias Hospitalarias, Complejo Hospitalario Universitario de A Coruña (CHUAC), Sistema Galego de Saúde (SERGAS), 15009 A Coruña, Spain 66Grupo de Investigación en Interacciones Gen-Ambiente y Salud (GIIGAS) - Instituto de Biomedicina (IBIOMED), Universidad de León, 24071 León, Spain 67Hospital Universitario de Getafe, Servicio de Genética, 28905 Madrid, Spain 68Ministerio de Salud Ciudad de Buenos Aires, Buenos Aires C1425EFD CABA, Argentina 69Hospital Universitario Virgen de las Nieves, Servicio de Análisis Clínicos e Inmunología, 18014 Granada, Spain 70IIS La Fe, Plataforma de Farmacogenética, 46026 Valencia, Spain 71Universidad de Valencia, Departamento de Farmacología, 46010 Valencia, Spain 72Data Analysis Department, Instituto de Investigación Sanitaria-Fundación Jiménez Díaz University Hospital - Universidad Autónoma de Madrid (IIS-FJD, UAM), 28040 Madrid, Spain 73Universidad de los Andes, Facultad de Ciencias, Bogotá 111711, Colombia D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 Human Molecular Genetics, 2022, Vol. 00, No. 00 | 3 74SIGEN Alianza Universidad de los Andes - Fundación Santa Fe de Bogotá, Bogotá 111711, Clombia 75Hospital General de Segovia, Medicina Intensiva, 40002 Segovia, Spain 76Hospital Universitario 12 de Octubre, Department of Immunology, 28041 Madrid, Spain 77Instituto de Investigación Sanitaria Hospital 12 de Octubre (imas12), Transplant Immunology and Immunodeficiencies Group, 28041 Madrid, Spain 78Instituto de Biomedicina (IBIOMED), Universidad de León, 24071 León, Spain 79Fundación Santa Fe de Bogota, Departamento Patologia y Laboratorios, Bogotá 111711, Colombia 80Unidad de Genética y Genómica Islas Baleares, 07120 Islas Baleares, Spain 81Hospital Universitario Son Espases, Unidad de Diagnóstico Molecular y Genética Clínica, 07120 Islas Baleares, Spain 82Genomics of Complex Diseases Unit, Research Institute of Hospital de la Santa Creu i Sant Pau, IIB Sant Pau, 08041 Barcelona, Spain 83Faculdade de Medicina, Universidade de Brasília, Brasilia 70910-900, Brazil 84Programa de Pós-Graduação em Ciências Médicas, Universidade de Brasília, Brasilia 70910-900, Brazil 85Programa de Pós-Graduação em Ciências da Saúde, Universidade de Brasília, Brasilia 70910-900, Brazil 86Hospital El Bierzo, Unidad Cuidados Intensivos, 24404 León, Spain 87Hospital Universitario Mostoles, Medicina Interna, 28935 Madrid, Spain 88Universidad Francisco de Vitoria, 28223 Madrid, Spain 89Programa de Pós-Graduação em Biologia Animal, Universidade de Brasília, Brasília 70910-900, Brazil 90Programa de Pós-Graduação Profissional em Ensino de Biologia, Universidade de Brasília, Brasília 70910-900, Brazil 91Programa de Pós-Graduação Profissional em Ensino de Biologia (UnB), Universidade de Brasília, Brasília 70910-900, Brazil 92Programa de Pós-Graduação em Ciências Médicas, Universidade de Brasília, Brasília 70910-900, Brazil 93Universidad Complutense de Madrid, Department of Immunology, Ophthalmology and ENT, 28040 Madrid, Spain 94Universidade Federal do Pará, Núcleo de Pesquisas em Oncologia, Belém, Pará 66075-110, Brazil 95Infectious Diseases, Microbiota and Metabolism Unit, Center for Biomedical Research of La Rioja (CIBIR), 26006 Logroño, Spain 96Inditex, 15141 A Coruña, Spain 97GENYCA, 28220 Madrid, Spain 98Instituto Mexicano del Seguro Social (IMSS), Centro Médico Nacional Siglo XXI, Unidad de Investigación Médica en Enfermedades Infecciosas y Parasitarias, Mexico City 02990, Mexico 99Clinica Comfamiliar Risaralda, 660003 Pereira, Colombia 100Bellvitge Biomedical Research Institute (IDIBELL), Neurometabolic Diseases Laboratory, 08908 L’Hospitalet de Llobregat, Spain 101Catalan Institution of Research and Advanced Studies (ICREA), 08010 Barcelona, Spain 102Hospital Ophir Loyola, Departamento de Ensino e Pesquisa, Belém, Pará 66063-240, Brazil 103Unidad de Cuidados Intensivos, Hospital Clínico Universitario de Santiago (CHUS), Sistema Galego de Saúde (SERGAS), 15706 Santiago de Compostela, Spain 104Department of Preventive Medicine and Public Health, School of Medicine, Universidad Autónoma de Madrid, 28049 Madrid, Spain 105IdiPaz (Instituto de Investigación Sanitaria Hospital Universitario La Paz), 28046 Madrid, Spain 106IMDEA-Food Institute, CEI UAM+CSIC, 28049 Madrid, Spain 107Complejo Asistencial Universitario de León, 24071 León, Spain 108Instituto de Investigación Biosanitaria de Granada (ibs GRANADA), 18012 Granada, Spain 109Universidad de Granada, Departamento Bioquímica, Biología Molecular e Inmunología III, 18071 Granada, Spain 110Hospital Infanta Elena, Allergy Unit, Valdemoro, 28342 Madrid, Spain 111Instituto de Investigación Sanitaria-Fundación Jiménez Díaz University Hospital - Universidad Autónoma de Madrid (IIS-FJD, UAM), 28040 Madrid, Spain 112Faculty of Medicine, Universidad Francisco de Vitoria, 28223 Madrid, Spain 113Hospital Universitario Infanta Leonor, 28031 Madrid, Spain 114Complutense University of Madrid, 28040 Madrid, Spain 115Gregorio Marañón Health Research Institute (IiSGM), 28007 Madrid, Spain 116Haemostasis and Thrombosis Unit, Hospital de la Santa Creu i Sant Pau, IIB Sant Pau, 08041 Barcelona, Spain 117Hospital Clinico Universitario de Valladolid, Servicio de Anestesiologia y Reanimación, 47003 Valladolid, Spain 118Universidad de Valladolid, Departamento de Cirugía, 47005 Valladolid, Spain 119Hospital Clinico Universitario de Valladolid, Servicio de Hematologia y Hemoterapia, 47003 Valladolid, Spain 120Hospital de Niños Ricardo Gutierrez, Buenos Aires C1425EFD CABA, Argentina 121Fundación Universitaria de Ciencias de la Salud, 113827 Bogotá, Colombia 122Spanish National Cancer Research Centre, Familial Cancer Clinical Unit, 28029 Madrid, Spain 123University Hospital of Burgos, 09006 Burgos, Spain 124Universidad Simón Bolívar, Facultad de Ciencias de la Salud, 080002 Barranquilla, Colombia 125Centro para el Desarrollo de la Investigación Científica, 1255 Asunción, Paraguay 126Institute for Molecular Medicine Finland (FIMM), 00014 Univerisity of Helsinki, Finland 127McGill University, Department of Human Genetics, H3A 0G4 Montréal, Québec, Canada 128Lady Davis Institute, Jewish General Hospital, McGill University, H3T 1E2 Montréal, Québec, Canada 129Kyoto-McGill International Collaborative School in Genomic Medicine, Graduate School of Medicine, Kyoto University, 606-8501 Kyoto, Japan 130Research Fellow, Japan Society for the Promotion of Science, 102-0083 Tokyo, Japan 131University of Milano-Bicocca, 20126 Milano, Italy 132Institute for Molecular Medicine Finland, Univerisity of Helsinki, 00014 Helsinki, Finland 133Institute of Clinical Molecular Biology, Christian-Albrechts-University, 24118 Kiel, Germany 134University Hospital Schleswig-Holstein, Campus Kiel, 24118 Kiel, Germany 135Department of Epidemiology, Biostatistics and Occupational Health, McGill University, H3A 0G4 Montréal, Québec, Canada 136Digestive Diseases Unit, Virgen del Rocio University Hospital, Institute of Biomedicine of Seville, University of Seville, 41103 Seville, Spain 137Centre for Biomedical Network Research on Hepatic and Digestive Diseases (CIBEREHD), Instituto de Salud Carlos III, 28029 Madrid, Spain 138Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute - Donostia University Hospital, University of the Basque Country (UPV/EHU), 20014 San Sebastian, Spain. 139Centre de Génétique Humaine, Hôpital Erasme, Université Libre de Bruxelles, 1070 Brussels, Belgium 140Liver Unit, Department of Internal Medicine, Hospital Universitari Vall d’Hebron, Vall d’Hebron Barcelona Hospital Campus, 08035 Barcelona, Spain 141Universitat Autònoma de Barcelona, Departament de Medicina, Bellatera, 08193 Barcelona, Spain 142Vall d’Hebron Institut de Recerca (VHIR), Liver Diseases, 08035 Barcelona, Spain 143Novo Nordisk Foundation Center for Protein Research, Disease Systems Biology, Faculty of Health and Medical Sciences, University of Copenhagen, DK-2200 Copenhagen, Denmark 144GENYO, Centre for Genomics and Oncological Research: Pfizer, University of Granada, Andalusian Regional Government, 18016 Granada, Spain 145Institute of Human Genetics, University Hospital Bonn, Medical Faculty University of Bonn, 53127 Bonn, Germany 146Genetica Medica, Azienda Ospedaliero-Universitaria Senese, 53100 Siena, Italy D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 4 | Human Molecular Genetics, 2022, Vol. 00, No. 00 147Universitat Autònoma de Barcelona, Departament de Medicina, Bellatera, 08193 Barcelona, Spain 148University of Siena, Medical Genetics, 53100 Siena, Italy 149Azienda Ospedaliero-Universitaria Senese, Genetica Medica, 53100 Siena, Italy 150, Med Biotech Hub and Competence Center, Department of Medical Biotechnologies, University of Siena, 53100 Siena, Italy 151Technical University of Munich, School of Medicine, University Hospital rechts der Isar, Department of Internal Medicine II, 80333 Munich, Germany 152Department of Transfusion Medicine and Hematology, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Università degli Studi di Milano, 20126 Milano, Italy 153Mucosal Immunology Lab, Unidad de Excelencia del Instituto de Biomedicina y Genética Molecular (IBGM, Universidad de Valladolid-CSIC), 47005 Valladolid, Spain 154Hospital Universitario Clinico San Cecilio, 18016 Granada, Spain 155Instituto de Investigación Ibs, Granada, 18012 Granada, Spain 156University of Liege. GIGA-Insitute, B- 4000 Liege, Belgium 157Liege University Hospital (CHU of Liege), B- 4000 Liege, Belgium 158Biomedical Research Institute Sant Pau (IIB Sant Pau), Stroke Pharmacogenomics and Genetics Group, 08041 Barcelona, Spain 159Oslo University Hospital, Department of Microbiology, 0424 Oslo, Norway 160Institute of Clinical Medicine, University of Oslo, 0424 Oslo, Norway 161Department of Liver and Gastrointestinal Diseases, Biodonostia Health Research Institute - Donostia University Hospital, University of the Basque Country (UPV/EHU), Ikerbasque, 20014 San Sebastian, Spain 162Department of Surgical Sciences, Anaesthesiology and Intensive Care Medicine, Uppsala University, 751 05 Uppsala, Sweden 163Humanitas University, Department of Biomedical Sciences, 20089 Milan, Italy 164IRCCS Humanitas Research Hospital, Rozzano, 20089 Milan, Italy 165Heart Institute (InCor)/University of Sao Paulo Medical School, 05508-070 Sao Paulo, Brazil 166Osakidetza Basque Health Service, Donostialdea Integrated Health Organisation, Clinical Biochemistry Department, 20006 San Sebastian, Spain 167Norwegian PSC Research Center and Section of Gastroenterology, Dept Transplantation Medicine, Oslo University Hospital, 0424 Oslo, Norway 168Research Institute of Internal Medicine, Oslo University Hospital, 0424 Oslo, Norway 169Fonds de la Recherche Scientifique (FNRS), B – 1000 Brussels 170Institute for Biomedical Research of Barcelona (IIBB), National Spanish Research Council (CSIC), 08028 Barcelona, Spain 171Institut d’Investigacions Biomediques August Pi i Sunyer (IDIBAPS), 08036 Barcelona, Spain 172Institute for Environmental Medicine, Karolinska Institutet, 171 65 Solna, Sweden 173Institute of Virology, Technical University Munich/Helmholtz Zentrum München, D-85764 Munich, Germany 174Institute of Psychiatric Phenomics and Genomics, University Hospital, LMU Munich University, 80539 Munich, Germany 175Department of Psychiatry, University Hospital, LMU Munich University, 80539 Munich, Germany 176Research Institute of Internal Medicine, Oslo University Hospital, 0318 Oslo, Norway 177Università degli Studi di Milano, Department of Pathopgysiology and Transplantation, 20126 Milano, Italy 178Department of Transfusion Medicine and Hematology, Precision Medicine, Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, 20122 Milano, Italy 179Karolinska Institutet, Department of Neuroscience, 171 77 Stockholm, Sweden 180Max-Planck Institute for Evolutionary Anthropology, 04103 Leipzig, Germany 181Department of Human Genetics, Epidemiology, Biostatistics and Occupational Health, McGill University, H3A 0G4 Montréal, Québec, Canada 182King’s College London, Department of Twin Research, London, WC2R 2LS, United Kingdom 183Massachusetts General Hospital, Harvard Medical School, Boston, MA 02114, USA 184Research Center and Memory clinic, ACE Alzheimer Center Barcelona, Universitat Internacional de Catalunya, 08028 Barcelona, Spain 185Centre for Biomedical Network Research on Neurodegenerative Diseases (CIBERNED), Instituto de Salud Carlos III, 28029 Madrid, Spain 186CIEN Foundation/Queen Sofia Foundation Alzheimer Center, 28031 Madrid, Spain 187Hospital Universitario de Valme, Unidad Clínica de Enfermedades Infecciosas y Microbiología, 41014 Sevilla, Spain 188Sección Genética Médica - Servicio de Pediatría, Hospital Clínico Universitario Virgen de la Arrixaca, Servicio Murciano de Salud, 30120 Murcia, Spain 189Departamento Cirugía, Pediatría, Obstetricia y Ginecología, Facultad de Medicina, Universidad de Murcia (UMU), 30100 Murcia, Spain 190Grupo Clínico Vinculado, Centre for Biomedical Network Research on Rare Diseases (CIBERER), Instituto de Salud Carlos III, 28029 Madrid, Spain 191Tecnologico de Monterrey, Escuela de Medicina y Ciencias de la Salud, 64718 Monterrey, Mexico 192Research Unit, Hospital Universitario N.S. de Candelaria, 38010 Santa Cruz de Tenerife, Spain 193Centre for Biomedical Network Research on Respiratory Diseases (CIBERES), Instituto de Salud Carlos III, 28029 Madrid, Spain 194Facultad de Ciencias de la Salud, Universidad Fernando Pessoa Canarias, 35450 Las Palmas de Gran Canaria, Spain *To whom correspondence should be addressed at: Unidad de Investigación, Hospital Universitario Nuestra Señora de Candelaria, Carretera del Rosario s/n, 38 010 Santa Cruz de Tenerife, Spain. Tel: +34 922602938; Fax: +34 922600545; Email: cflores@ull.edu.es (Carlos Flores); Centro de Investigación Biomédica en Red de Enfermedades Raras (CIBERER), ISCIII, Paseo de la Castellana 261, Madrid 28 046, Spain. Tel: +34 917277217; Fax: +34 912071030; Email: pablo.lapunzina@salud.madrid.org (Pablo Lapunzina); Center for Research in Molecular Medicine and Chronic Diseases (CIMUS), Av. Barcelona s/n, 15 782 Santiago de Compostela (A Coruña), Spain. Tel: +34 981951490; Fax: +34 881815403; Email: angel.carracedo@usc.es (Ángel Carracedo) †These authors contributed equally: Raquel Cruz, Silvia Diz-de Almeida. ‡These authors contributed equally: Carlos Flores, Pablo Lapunzina, Angel Carracedo. Abstract Here, we describe the results of a genome-wide study conducted in 11939 coronavirus disease 2019 (COVID-19) positive cases with an extensive clinical information that were recruited from 34 hospitals across Spain (SCOURGE consortium). In sex-disaggregated genome-wide association studies for COVID-19 hospitalization, genome-wide significance (P< 5×10−8) was crossed for variants in 3p21.31 and 21q22.11 loci only among males (P= 1.3×10−22 and P= 8.1×10−12, respectively), and for variants in 9q21.32 near TLE1 only among females (P= 4.4×10−8). In a second phase, results were combined with an independent Spanish cohort (1598 COVID-19 cases and 1068 population controls), revealing in the overall analysis two novel risk loci in 9p13.3 and 19q13.12, with fine-mapping prioritized variants functionally associated with AQP3 (P= 2.7×10−8) and ARHGAP33 (P= 1.3×10−8), respectively. The meta-analysis of both phases with four European studies stratified by sex from the Host Genetics Initiative (HGI) confirmed the association of the 3p21.31 and 21q22.11 loci predominantly in males and replicated a recently reported variant in 11p13 (ELF5, P= 4.1×10−8). Six of the COVID-19 HGI discovered loci were replicated and an HGI-based genetic risk score predicted the severity strata in SCOURGE. We also found more SNP-heritability and larger heritability differences by age (<60 or ≥60 years) among males than among females. Parallel genome-wide screening of inbreeding depression in SCOURGE also showed an effect of homozygosity in COVID-19 hospitalization and severity and this effect was stronger among older males. In summary, new candidate genes for COVID-19 severity and evidence supporting genetic disparities among sexes are provided. D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 Human Molecular Genetics, 2022, Vol. 00, No. 00 | 5 Introduction Coronavirus disease 2019 (COVID-19)—caused by the severe acute respiratory syndrome coronavirus 2 (SARS- CoV-2)—develops with wide clinical variability, ranging from asymptomatic infection to a life-threatening condition (1). Advanced age and the presence of comor- bidities are well-known major risk factors of COVID- 19 severity (2,3). In addition, male sex is another risk factor associated with COVID-19 severity, regardless of comorbidities (4). International genetic studies based on genome-wide association studies (GWAS) and/or comparative genome sequencing analyses have been conducted to identify genetic variants associated with COVID-19 severity (5,6). These studies have revealed the role of genes of the type-I interferon (IFN) signaling pathway as key players underlying disease severity (7–9). Besides, they have also identified other potential loci previously linked to lung function, respiratory diseases and autoimmunity (9). Regarding COVID-19 severity in males, sex-disaggregated genetic analyses have received limited attention despite the importance of sex disparities in clinical severity (10). Early studies suggested immune deficits presumably because of pre-existing neutralizing autoantibodies against type-I IFN in older male patients (11). The effects of autozygosity,measured as the change of the mean value of a complex trait because of inbreed- ing, have been useful to identify alternative genetic risk explanations and effects that traditionally are not cap- tured by GWAS (12). By analyzing the contribution of the inbreeding depression (ID) through the lens of the runs of homozygosity (ROH: genomic tracts where homozygous markers occur in an uninterrupted sequence), it is pos- sible to assess the importance of directional dominance or overdominance in the genetic architecture of com- plex traits (13). Even though this is a relatively modern approach, different studies have shown the importance of homozygosity in a large range of complex phenotypes, including anthropometric, cardiometabolic and mental traits (14–16). Through diverse nested sub-studies, the Spanish Coalition to Unlock Research on Host Genetics on COVID-19 (SCOURGE) consortium was launched in May 2020 aiming to find biomarkers of evolution and prognosis that can have an immediate impact on clinical management and therapeutic decisions in SARS-CoV- 2 infections. This consortium has recruited patients from hospitals across Spain and Latin America in close collaboration with the STOP-Coronavirus initiative (https://www.scourge-covid.org). Here, we describe the results of the first SCOURGE genome-wide studies of COVID-19 conducted in patients recruited in Spain. This dataset has not been used in any previous GWAS of COVID-19 that has been published to date. To the best of our knowledge, this is the first time that the impact of homozygosity is considered in COVID-19 studies, serving as a complement to the traditional GWAS to assess both the additive and dominant components of the genetic architecture of COVID-19 severity. Likewise, the ID analysis could also help to explain the strong effect of age in COVID-19 severity. Results Discovery phase In the SCOURGE study, 11 939 COVID-19 positive cases were recruited from 34 centers (Supplementary Mate- rial, Table S1) between March and December 2020. All diagnosed cases were classified in a five-level sever- ity scale (Table 1). Two untested Spanish sample col- lections were used as general population controls in some analyses: 3437 samples from the Spanish DNA biobank (https://www.bancoadn.org) and 2506 samples from the GR@CE consortium (17). The discovery phase samples were genotyped with the Axiom Spain Biobank Array (Thermo Fisher Scientific). Details of quality con- trol (QC), ancestry inference and imputation are shown in the Materials and Methods section. Individuals with inferred European ancestry were used for association testing. After post-imputation filtering, 15 045 individ- uals (9371 COVID-19 positive cases and 5674 popula- tion controls) and 8933154 genetic markers remained in the SCOURGE European study (discovery). Clinical and demographic characteristics of European patients from SCOURGE included in the analysis are shown in Table 2. Population controls were 46.3% females with a mean age of 55.5 years (standard deviation, SD=16.2) and 53.7% males, with a mean age of 51 years (SD=13.04). The discovery phase of the GWAS was carried out with infection susceptibility and three severity outcomes (hospitalization, severe illness and critical illness), which were tested using three different control definitions (see Supplementary Material, Table S2). • A1 analysis: COVID-19 positive not satisfying the case condition and control samples from the general pop- ulation (COVID-19 untested). • A2 analysis: control samples from the general popu- lation. • C analysis: COVID-19 positive not satisfying the case condition. The GWAS was carried on by fitting logistic mixed regression models adjusting for age, sex and the first 10 principal components (PCs; see Materials and Methods). Summary statistics can be accessed from https://github. com/CIBERER/Scourge-COVID19. The SCOURGE Board of Directors has agreed to aggregate the GWAS summaries with those from the COVID-19 Host Genetics Initiative (HGI) in the data freeze 7 that has not been used for any published article to date. Supplementary Material, Table S3 shows the independent significant associated loci for hospitalization, severity, critical illness and infec- tion susceptibility risk, for global and sex-stratified anal- ysis in the SCOURGE dataset. However, considering the overlap between the findings for these analyses, only the main results for the A1 analysis are presented. D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 6 | Human Molecular Genetics, 2022, Vol. 00, No. 00 Table 1. Five-level severity scale used to classify SCOURGE patients Level Clinical findings Severity 0 (asymptomatic) Asymptomatic Severity 1 (mild) With symptoms, but without pulmonary infiltrates or need of oxygen therapy Severity 2 (moderate) With pulmonary infiltrates affecting <50% of the lungs or need of supplemental oxygen therapy Severity 3 (severe) Hospitalized with any of the following criteria: • PaO2 <65 mmHg or SaO2 <90% • PaO2/FiO2 < 300 • SaO2/FiO2 <440 • Dyspnea • Respiratory frequency≥22 bpm • Infiltrates affecting> 50% of the lungs Severity 4 (critical) Admission to the ICU or need of mechanical ventilation (invasive or non-invasive) Note: PaO2, partial pressure of oxygen in arterial blood; SaO2, saturation of oxygen in arterial blood; FiO2, fraction of inspired oxygen. Table 2. Baseline characteristics of European patients from SCOURGE included in the analysis Variable Global N=9371 Males N=4343 Females N=5028 Age—mean years (SD) 62.6 (17.9) 64.3 (16.3) 61.1 (19.1) Severity—N (%) 0—asymptomatic 582 (6.6) 161 (3.9) 421 (8.9) 1—mild 2689 (30.3) 748 (18.2) 1941 (40.8) 2—intermediate 2099 (23.6) 1093 (26.5) 1006 (21.1) 3—severe 2379 (26.8) 1300 (31.6) 1079 (22.7) 4—critical illness 1128 (12.7) 817 (19.8) 311 (6.5) Hospitalization—N (%) 5966 (63.8) 3436 (79.3) 2530 (50.5) Severe COVID-19—N (%) 3507 (39.2) 2117 (51.2) 1390 (28.9) Critical illness—N (%) 1128 (12.6) 817 (19.8) 311 (6.5) Comorbidities—N (%) Vascular/endocrinological 4099 (43.7) 2207 (50.8) 1892 (37.6) Cardiac 1057 (11.3) 634 (14.6) 423 (8.4) Nervous 773 (8.3) 341 (7.9) 432 (8.6) Digestive 264 (2.8) 153 (3.5) 111 (2.2) Onco-hematological 647 (6.9) 411 (9.5) 236 (4.7) Respiratory 905 (9.7) 565 (13.0) 340 (6.8) All analyses support the association of two known loci, i.e. 3p21.31 and 21q22.11. However, other suggestive associations were also found (Fig. 1 and Supplementary Material, Fig. S1). Strikingly, the leading signals found in the global (sex-aggregated) analysis were genome-wide significant in the analyses among males but not among females. Association in 3p21.31 was also found in the C analyses (rs10490770, P=3.8× 10−12) and once again, association was genome-wide significant only among males (males: P=3.9× 10−9, females: P=4.6×10−5). However, the leading variant of 9q21.32 (near TLE1 gene) reached genome-wide significance among females only (similarly, in the C analysis for females, rs140152223, P=2.11×10−6). Several variants (rs17763742 near LZTFL1, rs2834164 in IFNAR2 and rs1826292621 near TLE1) showed a significant difference in effect sizes (SNP∗sex interaction P<0.0031, adjusted probability for 16 com- parisons) linked not only to hospitalization, but also to critical illness and infection risk. The A2 and C analyses did not reveal any additional significant loci (SupplementaryMaterial, Fig. S2).Although fine-mapping studies in 3p21.31 and 21q22.11 have led to gene and variant prioritization within these loci (Supplementary Material, Fig. S3), a Bayesian fine-mapping on the 9q21.32 did not allow to prioritize variants by their role as expression quantitative trait loci (eQTLs) or anticipate the function (Fig. 2). To assess if a higher prevalence of comorbidities in males may underlie these differential findings, we performed an additional C analysis in which the presence of comorbidities was tested for association within hospitalized patients. No significant association was found in either males or females (Supplementary Material, Fig. S4). Further explorations of the genetic associations with comorbidities are presented in the Supplementary Note. This GWAS phasewas also performed disaggregated by age (<60/≥60 years old), and by age and sex simultane- ously. Differences in effect sizes between both age groups were tested for the SNPs shown in the Supplementary Material, Table S3, in global and sex-specific analysis (Supplementary Material, Table S4). Significant findings were only found in the subgroup of males with <60 years old. This was also found in the C analysis for hospital- ization where association in 3p21.31 was significant only D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 Human Molecular Genetics, 2022, Vol. 00, No. 00 | 7 Figure 1.Association results of SCOURGE for global and sex-disaggregated A1 hospitalization analysis. (A) Manhattan plot of results from global analysis. A quantile–quantile plot of the global analysis is also shown as an inset. (B) Miami plot of results from sex-disaggregated analyses (top:males and bottom: females). Figure 2. Regional plot of a novel association at 9q21.32 found among females from the SCOURGE study. The x axis reflects the chromosomal position, and the y axis the −log(P-value). The sentinel variant is indicated by a diamond and all other variants are colour coded by their degree of LD with the sentinel variant in Europeans. Credible set for this signal is shown within a dashed square. The horizontal dotted blue line corresponds to the threshold for genome-wide significance (P=5× 10−8). D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 8 | Human Molecular Genetics, 2022, Vol. 00, No. 00 in males <60 years old (P=3.32× 10−9). Differences in effect size (significant age-interaction) were significant at 3p21.31 for severity and critical illness, and suggestive in hospitalization. Lookup of previously found COVID-19 host risk factors in the SCOURGE study Known significant loci for COVID-19 severity in 3p21.31 (near SLC6A20 and LZTFL1) and 21q22.11 (in IFNAR2) were clearly replicated at genome-wide significance in this study, specifically for risk of infection, hospitaliza- tion and severity. Three other loci, in 9q34.2 (in ABO), 12q24.13 (in OAS1) and 19p13.2 (near RAVER1 and TYK2), did not reach the genome-wide significance threshold but they were significant after correcting for the 390 tests performed in a lookup (13 SNPs and 30 analyses, significance threshold P< 1.3× 10−4). In agreement with previous results, ABO was mainly associated with the risk of infection. However, other loci as those in 3q12.3 (near RPL24) and 19p13.3 (near DPP9), previously found associated with COVID-19 severity, were not replicated in the SCOURGE Europeans. The complete list of results for the list of COVID-19 HGI significant loci (9) is shown in Figure 3 and in the Supplementary Material, Table S5. Figure 3 also reinforces the clear sex differences found in this study. Genetic risk score and the COVID-19 severity scale We developed a genetic risk score (GRS) combining the 13 leading variants associated with infection risk, hospital- ization or severity in the meta-analysis performed by the COVID-19 HGI (9). This GRS predicted the severity scale in SCOURGE but supporting the differentiation in three classes: (i) controls/asymptomatic/mild cases; (ii) moder- ate and severe cases and (iii) critical cases. (Supplemen- tary Material, Fig. S5). Simultaneously disaggregating by age (<60/≥60 years old) and sex, we identify the three severity classes in the subgroup of males with <60 years old, supporting the importance of this group in the over- all findings (Supplementary Material, Fig. S5). Details of this analysis can be found in Supplementary Note. Second study phase and meta-analysis with the discovery Results for hospitalization risk were meta-analysed with a second Spanish cohort, the CNIO study (see Materials andMethods). This study was filtered following the same QC and imputation procedures. The final dataset of the CNIO study included 2446 European individuals (1378 COVID-19 positive cases and 1068 population controls) and 8895721 markers. Table 3 shows the results that were genome-wide significant either in global or sex-stratified meta- analysis with SCOURGE. Besides the widely replicated loci at 3p21.31 and 21q22.11, three additional signals were found: chr9:33426577:A:G (intergenic to AQP7 and AQP3), chr17:45422978:G:C (intronic to ARHGAP27) and Figure 3. Lookup of previously found COVID-19 host risk factors in the SCOURGE study. Heatmap illustrating the results in all analyses performed in this study (rows) for the 13 leading variants in the COVID- 19 HGI study (columns). Each box illustrates the top associated variant within the focal region. The color (gray to dark red) indicates the strength (significance level) of the association in SCOURGE. Note: In three cases (chr12: 112919388, chr19: 4719431 and chr21: 33242905), the imputed variants did not pass QC filters in SCOURGE and they were replaced by the nearest QC-ed imputed variant (chr12:112919404, chr19:4719822 and chr21:33241950, respectively). chr19:35687796:G:A (intergenic to UPK1A and ZBTB32). Bayesian fine-mapping around chr17:45422978:G:C failed to prioritize a credible set of variants, hindering functional links of the locus. Functional assessments of the prioritized variants by the Bayesian fine-mapping analysis in the other two regions supported that they were eQTLs of the AQP3 and ARGAP33 genes, including whole blood and lung tissues (Fig. 4). D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 Human Molecular Genetics, 2022, Vol. 00, No. 00 | 9 Table 3. Genome-wide significant variants in global or sex-stratified meta-analysis between the SCOURGE and CNIO studies SNP chr:position Meta-ALL Meta-males Meta-females Nearest gene EA NEA beta SE P-value beta SE P-value beta SE P-value rs115679256 3:45587795 G A 0.43 0.08 1.1E−08 0.48 0.10 2.3E−06 0.40 0.11 2.9E−04 LIMD1 rs17763742 3:45805277 A G 0.60 0.05 4.1E−29 0.74 0.07 3.3E−25 0.43 0.08 4.5E−08 LZTFL1 rs35477280 3:45932600 G A 0.39 0.05 1.4E−17 0.48 0.06 6.3E−15 0.28 0.07 1.6E−05 FYCO1 rs4443214 3:46136372 T C 0.25 0.04 9.0E−09 0.26 0.06 1.4E−05 0.26 0.06 4.4E−05 XCR1 rs115102354 3:46180545 A G 0.41 0.07 1.6E−08 0.52 0.10 2.1E−07 0.32 0.10 2.0E−03 CCR3 rs10813976 9:33426577 A G 0.18 0.03 2.7E−08 0.16 0.04 2.5E−04 0.19 0.05 3.5E−05 AQP3 rs1230082 17:45422978 C G 0.16 0.03 2.1E−08 0.17 0.04 2.8E−05 −0.15 0.04 2.5E−04 ARHGAP27 rs77127536 19:35687796 G A −0.22 0.04 1.3E−08 −0.25 0.05 2.1E−06 −0.19 0.05 4.3E−04 UPK1A/ZTBT32 rs17860169 21:33240996 A G 0.19 0.03 2.3E−11 0.27 0.04 1.4E−11 0.12 0.04 3.7E−03 IFNAR2 Note: Representative SNPs were selected using the clump function of PLINK 1.9 (clumping parameters r2 = 0.5, Pval = 5× 10−8 and Pval2 = 0.05). EA, effect allele; NEA, non-effect allele; beta, effect coefficient; SE, standard error. Table 4. Results of European meta-analysis for hospitalization risk Meta-all Meta-males Meta-females SNP chr:position EA NEA beta SE P-value beta SE P-value beta SE P-value Nearest gene rs115679256 3:45587795 G A 0.37 0.06 1.3E−08 0.41 0.08 5.6E−07 0.36 0.09 1.6E−04 LIMD1 rs13078854 3:45820440 G A 0.53 0.04 6.7E−34 0.64 0.05 2.7E−33 0.38 0.06 1.0E−09 LZTFL1 rs41289622 3:45973053 T G 0.36 0.04 3.6E−21 0.44 0.05 3.4E−20 0.27 0.05 7.2E−07 FYCO1 rs115102354 3:46180545 A G 0.40 0.06 8.9E−12 0.48 0.07 6.8E−11 0.26 0.08 1.8E−03 XCR1 rs61882275 11:34482745 G A −0.12 0.02 1.0E−06 −0.17 0.03 4.1E−08 −0.08 0.03 1.3E−02 ELF5 rs4767028 12:112921383 A G −0.16 0.02 1.6E−10 −0.19 0.03 2.5E−09 −0.11 0.04 8.7E−04 OAS1 rs12609134 19:35687796 G A −0.19 0.03 2.3E−08 −0.22 0.04 9.5E−08 −0.13 0.05 6.0E−03 UPK1A/ZBTB32 rs17860169 21:33240996 A G 0.18 0.03 3.9E−12 0.21 0.03 1.6E−10 0.15 0.04 2.9E−05 IFNAR2 Note: Summary statistics of both phases (SCOURGE and CNIO) were meta-analysed with four additional sex-disaggregated European studies from the COVID-19 HGI consortium. EA, effect allele; NEA, non-effect allele; beta, effect coefficient; SE, standard error. These variants were also associated with the three severity groups previously outlined in SCOURGE by the GRS under a multinomial model (Supplementary Mate- rial, Table S6). Meta-analysis in independent European studies Hospitalization risk was meta-analysed with other Euro- pean studies combining both Spanish cohorts (SCOURGE and CNIO) with other four sex-disaggregated studies from the COVID-19 HGI consortium, namely: BelCOVID, GenCOVID, Hostage-Spain and Hostage-Italy (Table 4). Once again, the most outstanding significant loci were found at 3p21.31 and 21q22.11 (in global and male- specific analyses), and three additional loci reached genome-wide significance in the meta-analysis of males: chr12:11292383:A:G (in OAS1), chr19:35687796:G:A (inter- genic to UPK1A and ZBTB32) and chr11:34482745:G:A (in ELF5). The 3p21.31 variants reached genome-wide significance in females, although at significantly lower level than in males despite the similar sample sizes (Z=3.33, P= 5× 10−4). Significance of two interesting loci revealed in the Spanish studies was reduced in the meta-analysis with other European studies, although still showed suggestive associations: that of 9q21.32 near TLE1 previously found only in females (female meta-analysis P= 5.4×10−7), and that of 9p13.3 near AQP3 (global meta-analysis, P= 1.23× 10−7). Heritability of COVID-19 hospitalization In the hospitalization risk analysis, we found that common variants (minor allele frequency, MAF>1%) explain 27.1% (95% confidence interval, CI: 19.0–35.3%) of heritability on the observed scale (corresponding to 13.1% [95%CI: 9.2–17.0%] on the liability scale, assuming a prevalence of 0.5%; Fig. 5).We observed less heritability among females than males (2.9% [95%CI: 0.00–10.6%] in females and 17.0% [95%CI: 9.2–24.9%] in males on the liability scale). In agreement with observations suggesting an accumulation of non-genetic risk factors with age, especially among males (11,18), we observed larger heritability differences by age groups amongmales (40.2% [95%CI: 22.8–57.5%] in <60 years versus 17.6% [95%CI: 0.00–38.0%] in ≥60 years on the liability scale) than among females (9.1% [0.00–31.3%] in <60 years versus 13.7% [0.00–29.6%] in ≥60 years on the liability scale). ID and COVID-19 outcomes ROH calling was performed in the European QC-ed genotyped dataset. Inbreeding depression analyses are described in Materials and Methods section and Supplemental Note. The median genomic inbreeding coefficient, FROH, for the entire SCOURGE study was 0.0048 (IQR=0.004). No differences were detected between males (FROH =0.004, IQR=0.0035) and females (FROH =0.0056, IQR=0.0038), or D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 10 | Human Molecular Genetics, 2022, Vol. 00, No. 00 Figure 4.Regional plots of novel association signals found from themeta- analysis between the SCOURGE and CNIO studies. Regional plots of novel association signals found in 9p13.3 (A–C), 17q21.31 (D–F) and 19q13.12 (G–I). The x axes reflect the chromosomal position, and the y axes the −log(P-value) of the SCOURGE-CNIO meta-analysis. On A, D and G the sentinel variant is indicated by a diamond and all other variants are color coded by their degree of LD with the sentinel variant in Europeans. Whenever a concise set of variants was prioritized, a credible set is shown within a dashed square. The horizontal dotted blue line corresponds to the threshold for genome-wide significance (P=5×10−8). In the rest of panels, the x axes reflect the chromosomal position, and the y axes the −log(P-value) resulting from the eQTL analyses in whole blood (B, E and H) and in the lung (C, F and I) whenever a significant finding is available from GTEx v8. Figure 5. Forest plot of the SNP-heritability estimates for the COVID-19 hospitalization risk analysis on the liability scale. between younger and older individuals (FROH individuals< 60 years old = 0.004, IQR=0.0035; FROH individuals≥ 60 years old = 0.0052, IQR=0.0047, respectively; Supplementary Mate- rial, Fig. S6). Regarding the ID in COVID-19 outcomes, we detected a positive association of the FROH in COVID- 19 hospitalization risk (Fig. 6), severity risk and risk for critical illness (Supplementary Material, Table S7). Our results showed that the hospitalization odds for COVID- 19 patients with an FROH =0.0039 were 380% higher than individuals with FROH =0. No effect of the genomic relationship matrix (FGRM) was found. To assess whether ID in COVID-19 hospitalization in SCOURGE was different between sexes, we first tested the interaction between FROH and biological sex. FROH, sex and the interaction of both (FROH:Sex) were significant (FROH =4.7×10−3, Sex=1.0× 10−112, FROH:Sex=1.2× 10−3). This interaction was significant when comparing the hospitalized COVID-19 patients with different controls (A2 and C analyses, see Supplementary Material, Table S8). This interaction was also found significant with severity, but not with critical illness (Supplementary Material, Table S8). In sex-disaggregated analyses, we observed a sex-specific effect of inbreeding. FROH was significant in hospitalized males but not in females (Fig. 6 and Supplementary Material, Table S8). This sex- specific effect was also observed with severity and for critical illness (Supplementary Material, Table S8). We then assessed whether ID in COVID-19 hospitalization was different by age. We detected a significant inter- action between age and FROH for the three outcomes considered (hospitalization, severity and critical illness) (Supplementary Material, Table S9). After disaggregating SCOURGE by sex and age (<60, ≥60), we found that the ID for hospitalization and severity were signifi- cant mainly in older males (Fig. 6 and Supplemen- tary Material, Table S9). We detected significant ID for hospitalization and severity in males≥ 60 years old, but it was marginally significant in females (Fig. 6 and D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 Human Molecular Genetics, 2022, Vol. 00, No. 00 | 11 Figure 6. Effect of the ID on COVID-19 hospitalization in the SCOURGE cohort. Forest plots are shown for global analyses as well as for sex and age- disaggregated analyses. Supplementary Material, Table S9). Age and sex-specific effects in hospitalization and severity were robust across different experimental designs using different control groups (Supplementary Material, Fig. S7). Finally, we then aimed to replicate the ID results with hospitalization, assessing sex and age-specific effects, in a 4418 case-enriched European cohortmade of 16 studies from nine countries. Median FROH in this other European cohort was slightly higher than that of SCOURGE, 0.05 (0.009–0.0577). A positive and significant ID in COVID- 19 hospitalization was detected in this other European cohort when the entire cohort was considered (FROH Beta=18.22, P=3.33× 10−3). FGRM was not significant (FGRM Beta=−7.34, P=0.240). ID was also detected in hospitalized COVID-19 males but not in females (Male FROH Beta=16.22, P=3.31× 10−3; Female FROH Beta=15.65, P=0.269). FGRM was not significant in males or in female analyses.When disaggregating by age, it was possible to detect significant ID in hospitalization only in males ≥60 years old (FROH Beta=36.16, P=3.34×10−3) (Supplementary Material, Table S10). No evidencewas found ofmajor loci thatmay be exert- ing large effects. Rather, polygenicity seems to underlie the ID association. Different islands of ROH (ROHi) and regions of heterozygosity (RHZ) were found to be unique for hospitalized COVID-19 individuals (males and females) and non-hospitalized males, respectively (Supplementary Note, Supplementary Material, Table S11). Enrichment analysis of pathways based on the protein coding genes present in ROH islands were also different between sexes (Supplementary Note, Supplementary Material, Table S12), revealing links with coagulation and complement pathways in males. Discussion Here we report the replication of six COVID-19 loci across analyses and provide evidence supporting three additional loci, one of them specifically detected among females. Besides, our analyses provide new insights into disease severity disparities by sex and age and support the necessity of similarly stratified studies to increase the possibility of detecting additional risk variants. Our GWAS constitutes the largest study on COVID-19 genetic risk factors conducted in Spain, with an intrinsic design benefit that SCOURGE includes detailed clinical and genetic information gathered homogeneously across the country. Besides, the study included patients from the whole spectrum of COVID-19 severity covering from asymptomatic to life-threatening D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 12 | Human Molecular Genetics, 2022, Vol. 00, No. 00 disease. To date, most research on COVID-19 disease has focused on respiratory failure. However, the inclusion of a severity scale provided a unique opportunity to assess whether previously reported loci combined into a GRS model were associated with differential risk by strata. We warn, however, that the GRS model findings should be interpreted with caution as sex and age- differential results in some of the severity strata needs appropriate replication. Association was tested for four main variables: infection, hospitalization, severe illness and critical illness, and using different definitions of controls to align with the COVID-19 HGI. Irrespective of the tested outcomes or the definition of controls, the results were very similar, supporting the use of population controls to increase power in these studies and the utility of using hospitalization as a proxy of severity. However, our results from the GRS analysis reported a need to maintain separated categories for medium–severe and critical illness. We observed larger heritability differences by age groups among males than among females for COVID- 19 hospitalization, which have diverse support from the literature. On the one hand, there is robust evidence supporting that the presence of X-linked deleterious variants in the TLR7 gene are causal for life-threatening COVID-19 only affecting males (19–21). Of note, most of these severe COVID-19 male patients were younger than 60 years (21). On the other hand, autoantibodies impair- ing type-I interferon signaling, which are supported to be strong determinants of critical COVID-19 pneumonia, are preferentially found amongmales older than 65 years (11,18). Taken together, this reconciles with the idea that non-genetic factors involved in severe COVID-19 are expected among older males. We clearly replicated previously reported associations at 3p21.31 (near SLC6A20 and LZTFL1-FYCO1) (7,9,22,23) and 21q22.11 (in IFNAR2) (7,9), and other findings in ABO, OAS1, TYK2 and ARHGAP27. We also found a differential effect between males and females for SNPs in 3p21.31 and 21q22.11. Although in the meta-analysis with other European studies the leading variants of 3p21.31 reached genome-wide significance in females, there was still a difference in effect size that, considering its magnitude, should not be disregarded. It is important to remark that these association signals found in males were not associated with the presence of comorbidities (see Sup- plementary Material, Fig. S4). In fact, genetic effects were only found for younger males (<60 years old), consistent with other studies (24) and strongly supporting those comorbidities outweigh genetic effects in disease out- comes in the older patients. Some novel genome-wide significant signals were found in our study, one in chromosome 19q13.12 (intergenic to UPK1A and ZBTB32, and also linked to the transcriptional regulation of ARHGAP33), and another in chromosome 9p13.3 (intergenic to AQP7 and AQP3). Interestingly,we also found two sex-specific signals: ELF5 in males and TLE1 in females. ELF5 has been recently reported as a new locus associated with critical illness in Europeans (25). Variants of this locus reached genome- wide significance in our male meta-analysis of European cohorts (P= 4.1× 10−8). As regards of TLE1, this locus should be taken as speculative as the signal did not reach the standard genome-wide significance in the study. However, given that the meta-analysis involved a low number of studies (and the top marker was not imputed in one of them), this result should be taken with caution as further sex-specific studies will be needed to validate this finding. TLE1 encodes for the transducin-like enhancer of split 1, a co-repressor of other transcription factors and sig- naling pathways. Besides repressing the transcriptional activity of FOXA2 and of the Wnt signaling, TLE1 has been shown to negatively regulate NF-κB, which is fun- damental in controlling inflammation and the immune response. The deficiency of TLE1 activity in mice results in enhancement of the NF-κB-mediated inflammatory response in diverse tissues including the lung (26). Inter- estingly, TLE1 is one of the 332 high-confidence SARS- CoV-2 protein–human protein interactions identified so far (27). Taken together, SARS-CoV-2 would be directly targeting the innate immunity and inflammation signal- ing pathways by interfering with the NF-κB activity. Thus, it is not surprising that TLE1 is a top-ranking regulator of inflammation that allows to transcriptionally distin- guish mild from severe COVID-19 (28). In the 19q13.12 locus, the most biologically plausible genes are ARHGAP33 (also showing the best functional support based on the fine-mapping variants) and ZBTB32. ARHGAP33 is transcriptionally regulated by IRF1—a prominent antiviral effector and IFN-stimulated gene (29). It also harbors NF-κB binding site that modifies its expression in human lymphoblastoid cell lines by the presence of genetic variants in the binding site linked to many inflammatory and immune-related diseases including sepsis, and bacterial and viral infection (30). Its expression is also altered in human induced pluripotent stem cells-derived pancreatic cultures in response to SARS-CoV-2 infection (31). ARHGAP33 was identified in an unbiased genome-wide CRISPR-based knockout screen in human Huh7.5.1 hepatoma cells infected by coronaviruses including SARS-CoV-2 and further inter- actome studies (32). With respect to the transcription factor ZBTB32, it has been shown to impair antiviral immune responses by affecting cytokine production and the proliferation of natural killer and T cells, and the generation of memory cells (33). In single cell studies, transcripts of ZBTB32were enriched in T follicular helper cells and were also expressed at significantly higher levels in hospitalized COVID-19 patients (34). AQP3 is expressed most strongly in the kidney collecting duct, the gastrointestinal tract, large airways (in basal epithelial cells and the nasopharynx), skin and the urinary bladder; whereas AQP7 is expressed primarily in the testis, fat cells and, to a lesser extent in a subsegment of the kidney proximal tubule (35). In D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 Human Molecular Genetics, 2022, Vol. 00, No. 00 | 13 addition, AQP3 is upregulated in the lung tissues during viral or bacterial-induced diffuse alveolar damage (36). Based on this, in the fact that SARS-CoV-2 interacts with host proteins with the highest expression in lung tissues (27), and the functional evidence linking the fine- mapped variants with eQTLs in lung tissues, our data support AQP3 as the most likely 9p13.3 gene driving the association with COVID-19 hospitalization. Many patients develop acute respiratory distress syndrome (ARDS) during severe COVID-19 (37), and one of the hallmarks of ARDS is the increase of fluid volume (edema) in the airspaces of the lung because of an increase in the alveolo-capillarymembrane permeability. Some of the aquaporins, including AQP3, essentially function as water transport pores between the airways and the pulmonary capillaries (38), are key in lung fluid clearance and the formation of this lung edema as a consequence of the lung injury (35). In fact, the use of aquaporin modulators in lung inflammation and edema has been proposed for potential use for the treatment of COVID-19 respiratory comorbidity (39). We have also shown for the first time that COVID- 19 severity risk suffers from ID, where individuals with higher levels of homozygosity associate with higher risk of being hospitalized and of developing severe COVID- 19. Our results also suggested that autozygous rare recessive variants found in ROH across the genome, rather than homozygous common variants in strong LD, are underlying the ID. Furthermore, the ID is stronger in males than in females suffering from COVID-19 hospitalizations, especially in males≥ 60 years old. Although these results may be found counterintuitive with the rest of findings, they are supported by the mutation accumulation senescence theory. Under this theory, alleles with detrimental effects that act in late life are expected to accumulate and cause senescence, thus increasing the ID (40). We detected further sex-specific effects of homozygosity through ROHi. In hospitalized males, coagulation and complement pathways, which have been previously associated with severe COVID- 19 (41), were enriched among the protein coding genes located in ROHi, highlighting the role of homozygosity whereas the Lectin pathway of complement activation is reported to be involved in the response to SARS- CoV-2 infection (42–44). In hospitalized females, PI3K- Akt signaling genes were found to be enriched in ROH islands, whose networks are affected by a great variety of viruses (45). Given that the effect of the genetic variants in SARS- CoV-2 severity is clearer among males and previous evi- dence on this regard, we elucubrate on the role of andro- gens in COVID-19 severity. Androgenic hormones have been suggested to be responsible of the excess male mortality observed in COVID-19 patients (46), and several lines of evidence suggest that the androgen receptor (AR) pathway is involved in the severity of SARS-CoV-2 infection: (i) A higher mortality rate among men has been established (47); (ii) A substantial proportion of individuals, both males and females, hospitalized for severe COVID-19 have androgenetic alopecia [AGA; (47)] and (iii) Most of the genes on COVID-19 severity in this study have been identified in male-only analyses, and these genes have been shown to interact with the AR. The following lines of evidence suggest the AR pathway is a mechanism responsible for some identified genes- COVID-19 severity relationship: (i) FYCO1 is regulated by the AR (48), and binding sites between the sex hormone receptor AR and FYCO1 have been demonstrated (48,49); (ii) There is a cross-talk between the IFN pathways and the androgen signaling pathways (50), and androgens are regulated by IFNs in human prostate cells (51); (iii) TMPRSS2, another gene associated with COVID-19 sever- ity in other studies, is induced by androgens through a distal AR binding enhancer (52); (iv) AR induces the expression of chemokine receptors such as CCR1; (v) Variants of LZTFL1 gene are likely pathogenic for male reproductive system diseases (53) and (vi) Genetic poly- morphisms in the AR (long polyQ alleles ≥23) and higher testosterone levels in subjectswith AR long-polyQ appear to predispose some men to develop more severe dis- ease (54). Thus, it is not unexpected to find that antian- drogen treatments are under the focus as treatment options and prophylaxis of severe COVID-19 (47) and that randomized controlled clinical trials with bicalutamide (NCT04374279), degarelix (NCT04397718) and spirono- lactone (NCT04345887) are currently underway. Materials and Methods Recruitment of cases and phenotype definitions for the discovery phase In Spain, 11 939 COVID-19 positive cases were recruited as part of SCOURGE study from 34 centers in 25 cities between March and December 2020. The complete list of hospitals or research centers and the number of samples that each contributed to the study is shown in Sup- plementary Material, Table S1. Study samples and data were collected by the participating centers, through their respective biobanks after informed consent, with the approval of the respective Ethic and Scientific Commit- tees. Thewhole project was approved by the Galician Eth- ical Committee Ref 2020/197. All samples and data were processed following normalized procedures. Study data were collected and managed using REDCap electronic data capture tools hosted at Centro de Investigación Biomédica en Red [CIBER; (55,56); Supplementary Mate- rial, Supplemental Note]. Individuals were diagnosed as COVID-19 positive through a PCR-based test (81.7% of cases) or according to local clinical (3.4%) and laboratory procedures (antibody test: 14.2%; other microbiological tests: 0.7%). All cases were classified in a five-level sever- ity scale (Table 1). Two Spanish sample collectionswith unknownCOVID- 19 status were included as general population controls in some analyses: 3437 samples from the Spanish DNA biobank (https://www.bancoadn.org) and 2506 D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 14 | Human Molecular Genetics, 2022, Vol. 00, No. 00 samples from the GR@CE consortium (17). General population controls were collected from branches of the National Blood Service from adult unrelated individuals self-reporting Spanish origin and absence of personal and familial history of diseases including infectious, cancerous, blood and circulatory, endocrine, mental or behavioral, nervous, vision, hearing, respiratory, immunological, bone, congenital, skin and digestive. Second phase: the CNIO study A total of 1598 COVID-19 cases from six different Spanish Biobanks (Biobanco CNIO, Biobanco Vasco, Biobanco Hos- pital Ramón y Cajal, Biobanco Hospital Puerta de Hierro, Biobanco Hospital San Carlos, and Banco Nacional de ADN) were obtained according to the ethical committee approval CEI PI 34_2020-v2. In addition, 1068 individuals from Spanish DNA biobank with unknown COVID-19 status were included as healthy controls in the anal- ysis whenever necessary. Classification as healthy was based on self-reported absence of cardiovascular, renal, pulmonary, hepatic, hematological illnesses or any other chronic conditions, which require continuous treatment, hepatitis B, C infections or acquired immunodeficiency syndrome (AIDS). No clinical characterization was per- formed on any subject, no information from medical record was incorporated and no medical testing was per- formed on these individuals. We will refer to these cases and controls as the Centro Nacional de Investigaciones Oncológicas (CNIO) study. Genotyping The discovery phase samples were genotyped with the Axiom Spain Biobank Array (Thermo Fisher Scientific) following the manufacturer’s instructions in the San- tiago de Compostela Node of the National Genotyping Center (CeGen-ISCIII; http://www.usc.es/cegen). This array contains 757 836 markers, including rare variants selected in the Spanish population. Genomic DNA was obtained from peripheral blood and isolated using the Chemagic DNA Blood100 kit (PerkinElmer Chemagen Technologies GmbH), following the manufacturer’s recommendations. For the second phase study samples, a total of 250 ng of DNA was processed according to the Infinium HTS assay Protocol (Part # 15045738 Rev. A, Illumina), includ- ing amplification, fragmentation and hybridization using the Global Screening Array Multi-disease v3.0. This array contains a total of 730 059 markers and was scanned on an iScan platform (Illumina, Inc.). Clustering and genotype calling were performed using Genome Studio v2.0.4 (Illumina, Inc.). Quality control A QC procedure was carried out for the SCOURGE study samples and control datasets. First, a list of probe sets was removed based on poor cluster separation or exces- sive MAF difference from The 1000 Genomes Project data (1KGP) (57). Then, the following QC steps were applied using PLINK 1.9 (58) and a custom R script. We excluded variants with MAF< 1%, call rate<98%, a difference in missing rate between cases and controls >0.02, or devi- ating from Hardy–Weinberg equilibrium (HWE) expecta- tions [P< 1× 10−6 in controls, P< 1×10−10 in cases, with a mid-P adjustment (59)]. Samples with a call rate<98% and those in which heterozygosity rate deviated >5 SD from the mean heterozygosity of the study were also removed. To assess kinship and assign ancestries, autosomal SNPs (MAF> 5%) were pruned with PLINK using a window of 1000 markers, a step size of 80 and a r2 of 0.1. In addition, high-linkage disequilibrium (LD) regions described in Price et al. (60) were also excluded. A subset of 131 937 independent SNPs was used to evaluate kinship (IBD estimation) in PLINK. Given the possible confusion between relatedness and ancestry, we removed only one individual from each pair of individuals with PI_HAT> 0.25 (second-degree relatives) that showed a Z0, Z1 and Z2 coherent pattern (according to theoretical expected values for each relatedness level). The unrelated SCOURGE individuals were merged with samples from 1KGP and the common SNPs were LD-pruned as previously indicated. Ancestry was then inferred with Admixture (61) using the defined 1KGP superpopulations. Those individuals with an estimated probability >80% of pertaining to European ancestry were defined as European (N=15571). Genomic PCs were also computed using a LD-pruned (r2 < 0.1 with a window size of 1000 markers) subset of genotyped SNPs passing quality check for controlling the population structure in the GWAS. The CNIO study data was filtered following the same QC procedures, where 220 individuals were identified as non-European and, therefore, were excluded from fur- ther analysis. Variant imputation Imputation was conducted based on the TOPMed version r2 reference panel [GRCh38; (62)] in the TOPMed Impu- tation Server. After post-imputation filtering (Rsq>0.3, HWE P> 1× 10−6, MAF> 1%), 15 045 individuals (9371 COVID-19 positive cases and 5674 population controls) and 8933154 genetic markers remained in the SCOURGE European study (discovery). The final dataset of the CNIO study (replication) included 2446 individuals (1378 COVID-19 positive cases and 1068 population controls) and 8895721 markers. For directly genotyped variants, the original genotype was maintained in place of the imputed data. Statistical analysis Association testing was computed by fitting logistic mixed regression models adjusting for age, sex and the first 10 ancestry-specific PCs. SNPRelate (63) was used for prior LD-pruning and data management. Association analyses were performed in SAIGEgds (64), which implements the SAIGE (65) two-step mixed model D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 Human Molecular Genetics, 2022, Vol. 00, No. 00 | 15 methodology and the SPA test while using more efficient objects for genotype storage. A null model was fitted in the first step using the LD-pruned genotyped variants (MAF> 0.5%, missing rate< 98%) to estimate variance components and the genetic relationship matrix. Then, in a second step, association analyses were performed for both genotyped and imputed SNPs. Significance was established at P< 5× 10−8 after meta-analysis of the discovery and the second study phases. To align the results with those from the COVID-19 HGI, three outcomes were evaluated in relation to severity: hospitalization, severe COVID-19 (severity≥ 3) and very severe COVID-19 (severity = 4, critical illness). For each comparison, three control definitions (A1, A2 and C) were used (Supplementary Material, Table S2). In addition, the risk to COVID-19 infection was also analysed by comparing all COVID-19 positive cases with control samples from the general population. All analyseswere conducted for each complete dataset and stratified by sex and age (<60 years, ≥60 years). The SNP∗sex and SNP∗age-interaction terms were tested for each SNP in the subset of clumped signals, adjusting the models for the same covariates. Then, the main results of both Spanish cohorts (SCOURGE and CNIO) for the overall and sex-stratified analyses were meta-analysed assuming a fixed-effects model using METAL (66). Because of the similarity of both the SCOURGE and CNIO studies in the clinical variables recorded and,more importantly, in the definition of the severity scale, the leading variants from the significant and validated loci in the hospitalization analysis were also analysed under a multinomial model (Supplementary Material, Supple- mental Note). Meta-analysis in independent European studies In order to validate the findings in independent study samples of European ancestry, a meta-analysis of hos- pitalization risk was performed for the overall and sex- stratified summary statistics of both Spanish cohorts (SCOURGE and CNIO) and other four sex-stratified Euro- peans studies from the COVID-19 HGI consortium (Bel- COVID, GenCOVID, Hostage-Spain and Hostage-Italy). Bayesian fine-mapping of GWAS findings Credible sets were calculated for the GWAS loci to identify a subset of variants most likely containing the causal variant at 95% confidence level, assuming that there is a single causal variant and that it has been tested. We used corrcoverage for R (67) to calculate the posterior probabilities of the variant being causal for all variants with r2 > 0.1 with the leading SNP and within 1 Mb. Variants were added to the credible set until the sumof the posterior probabilities was≥0.95.VEP (https:// www.ensembl.org/info/docs/tools/vep/index.html) and the V2G aggregate scoring from Open Targets Genetics (https://genetics.opentargets.org) were used to annotate the most prominent biological effects of the variants in the credible sets. Genetic risk score A GRS was created for the SCOURGE cohort individuals and population controls using the list of SNPs associated with hospitalization, severity or risk in the meta-analysis performed by the COVID-19 HGI [see Supplementary Material, Table S2 in (9)] to appraise its prediction power of the severity scale in SCOURGE. Details of this analysis can be found in Supplementary Note. SNP-heritability of COVID-19 severity We relied on GCTA-GREML 1.93.2beta (68) to assess the heritability of severe COVID-19 symptoms among SCOURGE patients, excluding those with cryptic related- ness or withmissing genotypes above 0.5% and assuming a prevalence of COVID-19 hospitalization of 0.5%. This analysis considered all patients (modelling for age, sex, sex∗age and the 10 first PCs), and males and females separately (modelling for age and the 10 first PCs). We also partitioned the variance to assess the heritability changes among the patients <60 or ≥60 years old. We focused on the 547 206 autosomal variants with MAF>1% and missingness <0.5%. Assuming 0.5% of prevalence of severe COVID-19, and at least 1500 cases and 1500 controls per stratum, we estimate >97.9% power to detect a heritability >0.2. ROH calling The ROH segments longer than 300 Kb were called in SCOURGE using PLINK 1.9 in the European QC- ed genotyped dataset (before imputation) with the following parameters: homozyg-snp 30, homozyg-kb 300, homozyg-density 30, homozyg-window-sn 30, homozyg-gap 1000, homozyg-window-het 1, homozyg-window-missing 5 and homozyg-window-threshold 0.05. No LD pruning was performed. Calculating genomic inbreeding coefficients Different genomic inbreeding coefficients were calcu- lated (69): FROH measures the actual proportion of the autosomal genome that is autozygous above a specific threshold of minimum ROH length. FROH = ∑n i=1 ROH > 1.5 Mb 3 Gb FGRM is an alternative genomic inbreeding coefficient that was obtained using PLINK’s parameter -ibc (Fhat3). This coefficient is described by Yang et al. (68) where N is the number of SNPs, pi is the reference allele frequency of the ith SNP, and xi is the number of copies of the reference allele. The reference allele frequencies were site-specific D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 16 | Human Molecular Genetics, 2022, Vol. 00, No. 00 and included only variants with MAF> 0.05. FGRM = 1N n∑ i ( x2i − ( 1 + 2pi ) xi + 2p2i ) 2pi ( 1 − pi ) Testing and replicating the ID Inbreeding depression is defined as the change in the mean phenotypic value in a population because of inbreeding (12,13). The ID was modelled in SCOURGE by a multiple logistic regression. The covariables used in this study were sex, age and the first 10 PCs. The results were replicated in a cohort gathered by Nakanishi et al. (24). This consists of clinical and genomic data from 4418 individuals of European ancestry (3946 hospitalized COVID-19 cases and 422 controls): 2597 males (1072 males<60 years old, 1525 males≥ 60 years old) and 1821 females (808 females< 60 years old, 1013 females≥60 years old). The cohort was built by harmonizing individual-level data from 16 different studies (24). The FROH and FGRM coefficients were obtained following the procedure explained previously. The model described previously with the same covariables (age, sex and the first then PCs) was applied in this cohort. Genome-specific effects on COVID-19 severity and hospitalization were tested through ID in genomic win- dows, ROH islands (ROHi) and regions of heterozygosity (RHZ) (Supplementary Material, Supplemental Note). Supplementary Material Supplementary Material is available at HMGJ online. Acknowledgements R.L.-R. is granted by Cátedra de Medicina Genómica IIS- Fundación Jiménez Díaz-UAM, M.B. by Nextgeneration EU funds. M.C., P.M., M.A.J.S., A.F.R. are granted by the Miguel Servet Programme (CP17/00006, CP16/00116, CPII20CIII/0001, CPII20CIII/0001 respectively) and B.A. by the Juan Rodés Programme (JR17/00020), all of them from Instituto de Salud Carlos III, and cofunded by European Union (ERDF) ‘A way of making Europe’. The contribution of the Centro National de Genotipado (CEGEN), and Centro de Supercomputación de Galicia (CESGA) for funding this project by providing supercom- puting infrastructures, is also acknowledged. Authors are also particularly grateful for the supply of material and the collaboration of patients, health professionals from participating centers and biobanks. Namely Biobanc- Mur, and biobancs of the Complexo Hospitalario Univer- sitario de A Coruña, Complexo Hospitalario Universitario de Santiago, Hospital Clínico San Carlos, Hospital La Fe, Hospital Universitario Puerta de Hierro Majadahonda— Instituto de Investigación Sanitaria Puerta de Hierro— Segovia de Arana, Hospital Ramón y Cajal, IDIBGI, IdISBa, IIS Biocruces Bizkaia, IIS Galicia Sur. Also biobanks of the Sistema de Salud de Aragón, Sistema Sanitario Público de Andalucía, and Banco Nacional de ADN. SCOURGE Cohort Group members and affiliations, HOSTAGE Cohort Group and GR@ACE Cohort Group (Supplementary Material). Conflict of Interest statement: The authors declare no com- peting interests. Funding Instituto de Salud Carlos III (COV20_00622 to A.C., COV20/00792 to M.B., COV20_00181 to C.A., COV20_1144 to M.A.J.S., PI20/00876 to C.F.); European Union (ERDF) ‘A way of making Europe’. Fundación Amancio Ortega, Banco de Santander (to A.C.), Estrella de Levante S.A. and Colabora Mujer Association (to E.G.-N.) and Obra Social La Caixa (to R.B.); Agencia Estatal de Investigación (RTC-2017-6471-1 to C.F.), Cabildo Insular de Tenerife (CGIEU0000219140 ‘Apuestas científicas del ITER para colaborar en la lucha contra la COVID-19’ to C.F.) and Fundación Canaria Instituto de Investigación Sanitaria de Canarias (PIFIISC20/57 to C.F.). References 1. Tang, D., Komish, P. and Kang, R. (2020) The hallmarks of COVID-19 disease. PLoS Pathog., 16, e1008536. https://doi.org/10.1371/journal.ppat.1008536. 2. Goyal, P., Choi, J., Pinheiro, L., Schenck, E., Chen, R., Jabri, A., Satlin, M., Campion, R., Nahid, M., Ringel, J. et al. (2020) Clinical characteristics of Covid-19 in new York City. N. Engl. J. Med., 382, 2372–2374. 3. Richardson, S., Hirsch, J., Narasimhan, M., Crawford, J., McGinn, T., Davidson, K. and the Northwell COVID-19 Research Con- sortium (2020) Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City area. JAMA, 323, 2052–2059. 4. Vahidy, F., Pan, A., Ahnstedt, H., Munshi, Y., Choi, H., Tiruneh, Y., Nasir, K., Kash, B., Andrieni, J. and McCullough, L. (2021) Sex differences in susceptibility, severity, and outcomes of coronavirus disease 2019: cross-sectional analysis from a diverse US metropolitan area. PLoS One, 16, e0245556. https://doi.org/10.1371/journal.pone.0245556. 5. The COVID-19 Host Genetics Initiative (2020) The COVID-19 host genetics initiative, a global initiative to elucidate the role of host genetic factors in susceptibility and severity of the SARS-CoV-2 virus pandemic. Eur. J. Hum. Genet., 28, 715–718. 6. Casanova, C., Su, H. and COVID Human Genetic Effort (2020) A global effort to define the human genetics of protective immu- nity to SARS-CoV-2 infection. Cell, 181, 1194–1199. 7. Pairo-Castineira, E., Clohisey, S., Klaric, L., Bretherick, A., Rawlik, K., Pasko, D., Walker, S., Parkinson, N., Fourman, M., Russell, C. et al. (2021) Genetic mechanisms of critical illness in COVID-19. Nature, 591, 92–98. 8. Zhang, Q., Bastard, B., Liu, Z., Le Pen, J., Moncada-Velez, M., Chen, J., Ogishi, M., Sabli, I., Hodeib, S., Korol, C. et al. (2020) Inborn errors in type I IFN immunity in patients. Science, 370, eabd4570. https://doi.org/10.1126/science.abd4570. 9. COVID-19 Host Genetics Initiative (2021) Mapping the human genetic architecture of COVID-19. Nature, 600, 472–477. D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 Human Molecular Genetics, 2022, Vol. 00, No. 00 | 17 10. Brady, E., Nielsen, M., Andersen, J. and Oertelt-Prigione, S. (2021) Lack of consideration of sex and gender in COVID-19 clinical studies. Nat. Commun., 12, 4015. https://doi.org/10.1038/s41467-021-24265-8. 11. Bastard, P., Rosen, L., Zhang, Q., Michailidis, E., Hoffmann, H., Dorgham, Z., Philippot, Q., Rosain, J., Béziat, V., Manry, J. et al. (2020) Autoantibodies against type I IFNs in patients with life-threatening COVID-19. Science, 370, eabd4570. https://doi.org/10.1126/science.abd4585. 12. Charlesworth, D. and Willis, H. (2009) The genetics of inbreeding depression. Nat. Rev. Genet., 10, 783–796. 13. Ceballos, F., Joshi, P., Clark, D., Ramsay, M. and Wilson, J. (2018) Runs of homozygosity: windows into population history and trait architecture. Nat. Rev. Genet., 19, 220–234. 14. Ceballos, F., Hazelhurst, S., Clark, D., Agongo, G., Asiki, G., Boua, P., Gómez-Olivé, X., Mashinya, F., Norris, S., Wilson, J. et al. (2020) Autozygosity influences cardiometabolic disease- associated traits in the AWI-gen sub-Saharan African study.Nat. Commun., 11, 5754. https://doi.org/10.1038/s41467-020-19595-y. 15. Clark, D., Okada, Y., Moore, K., Mason, D., Pirastu, N., Gandin, I., Mattsson, H., Barnes, C., Lin, K., Zhao, J. et al. (2019) Associations of autozygosity with a broad range of human phenotypes. Nat. Commun., 10, 4957. https://doi.org/10.1038/s41467-019-12283-6. 16. Moreno-Grau, S., Fernández, M., de Rojas, I., Garcia-González, P., Hernández, I., Farias, F., Budde, J., Quintela, I., Madrid, L., González-Pérez, A. et al. (2021) Long runs of homozygosity are associated with Alzheimer’s disease. Transl. Psychiatry, 11, 142. https://doi.org/10.1038/s41398-020-01145-1. 17. Moreno-Grau, S., de Rojas, I., Hernández, I., Quintela, I., Mon- trreal, L., Alegret, M., Hernández-Olasagarre, B., Madrid, L., González-Perez, A., Maroñas, O. et al. (2019) Genome-wide asso- ciation analysis of dementia and its clinical endophenotypes reveal novel loci associated with Alzheimer’s disease and three causality networks: the GR@ACE project.Alzheimers Dement., 15, 1333–1347. 18. Bastard, P., Gervais, A., Le Voyer, T., Rosain, J., Philippot, Q., Manry, J., Michailidis, E., Hoffmann, H., Eto, S., Garcia-Prat, M. et al. (2021) Autoantibodies neutralizing type I IFNs are present in ∼4% of uninfected individuals over 70 years old and account for ∼20% of COVID-19 deaths. Sci. Immunol., 6, EABL4340. https://doi.org/10.1126/sciimmunol.abl4340. 19. van der Made, C., Simons, A., Schuurs-Hoeijmakers, J., van den Heuvel, G., Mantere, T., Kersten, S., van Deuren, R., Steehouwer, M., van Reijmersdal, S., Jaeger, M. et al. (2020) Presence of genetic variants among young men with severe COVID-19. JAMA, 324, 663–673. 20. Fallerini, C., Daga, S., Mantovani, S., Benetti, E., Picchiotti, N., Francisci, D., Paciosi, F., Schiaroli, E., Baldassarri, M., Fava, F. et al. (2021) Association of Toll-like receptor 7 vari- ants with life-threatening COVID-19 disease in males: find- ings from a nested case-control study. Elife, 10, e67569. https://doi.org/10.7554/eLife.67569. 21. Asano, T., Boisson, B., Onodi, F., Matuozzo, D., Moncada- Velez, M., Maglorius Renkilaraj, M., Zhang, P., Meertens, L., Bolze, A., Materna, M. et al. (2021) X-linked recessive TLR7 deficiency in ∼1% of men under 60 years old with life-threatening COVID-19. Sci. Immunol., 6, eabl4348. https://doi.org/10.1126/sciimmunol.abl4348. 22. Severe Covid-19 GWAS Group, Ellinghaus, D., Degenhardt, F., Bujanda, L., Buti, M., Albillos, A., Invernizzi, P., Fernández, J., Prati, D., Baselli, G. et al. (2020) Genomewide association study of severe Covid-19 with respiratory failure. N. Engl. J. Med., 383, 1522–1534. 23. Shelton, J., Shastri, A., Ye, C., Weldon, C., Filshtein-Sonmez, T., Coker, D., Symons, A., Esparza-Gordillo, J., 23andMe COVID-19 Team, Aslibekyan, S. et al. (2021) Trans-ancestry analysis reveals genetic and nongenetic associations with COVID-19 susceptibil- ity and severity. Nat. Genet., 53, 801–808. 24. Nakanishi, T., Pigazzini, S., Degenhardt, F., Cordioli, M., Butler- Laporte, G., Maya-Miles, D., Bujanda, L., Bouysran, Y., Niemi, M., Palom, A. et al. (2021) Age-dependent impact of the major com- mon genetic risk factor for COVID-19 on severity and mortality. J. Clin. Invest., 131, e152386. https://doi.org/10.1172/JCI152386. 25. Kousathanas, A., Pairo-Castineira, E., Rawlik, K., Stuckey, A., Odhams, C., Walker, S., Russell, C., Malinauskas, T., Mil- lar, J., Elliott, K. et al. (2022) Whole genome sequencing reveals host factors underlying critival Covid-19. Nature. https://doi.org/10.1038/s41586-022-04576-6. 26. Ramasamy, S., Saez, B., Mukhopadhyay, S., Ding, D., Ahmed, A., Chen, X., Pucci, F., Yamin, R., Wang, J., Pittet, M. et al. (2016) Tle1 tumor suppressor negatively regulates inflammation in vivo and modulates NF-κB inflammatory pathway. PNAS, 113, 1871–1876. 27. Gordon, D., Jang, G., Bouhaddou,M., Xu, J., Obernier, K.,White, K., O’Meara,M.,Rezelj,V.,Guo, J., Swaney,D. et al. (2020) A SARS-CoV- 2 protein interaction map reveals targets for drug repurposing. Nature, 583, 459–468. 28. de Jong, T., Guryev, V. and Moshkin, Y. (2021) Estimates of gene ensemble noise highlight critical pathways and predict disease severity in H1N1, COVID-19 and mortality in sepsis patients. Sci. Rep., 11, 10793. https://doi.org/10.1038/s41598-021-90192-9. 29. Schoggins, J. and Rice, C. (2011) Interferon-stimulated genes and their antiviral effector functions. Curr. Op. Virol., 1, 519–525. 30. Karczewski, K., Dudley, J., Kukurba, K., Chen, R., Butte, A., Mont- gomery, S. and Snyder, M. (2013) Systematic functional regu- latory assessment of disease-associated variants. PNAS, 110, 9607–9612. 31. Shaharuddin, S., Wang, V., Santos, R., Gross, A., Wang, Y., Jawanda, H., Zhang, Y., Hasan,W., Garcia, G., Jr., Arumugaswami, V. et al. (2021) Deleterious effects of SARS-CoV-2 infection on Human pancreatic cells. Front. Cell. Infect. Microbiol., 11, 678482. https://doi.org/10.3389/fcimb.2021.678482. 32. Wang, R., Simoneau, C., Kulsuptrakul, J., Bouhaddou, M., Trav- isano, K., Hayashi, J., Carlson-Stevermer, J., Zengel, J., Richards, C., Fozouni, P. et al. (2021) Genetic screens identify host factors for SARS-CoV-2 and common cold coronaviruses. Cell, 184, 106–119. 33. Shin, H., Kapoor, V., Kim, G., Li, P., Kim, H., Suresh, M., Kaech, S., Wherry, E., Selin, L., Leonard, W., Welsh, R.M. and Berg, L.J. (2017) Transient expression of ZBTB32 in anti-viral CD8+ T cells limits the magnitude of the effector response and the generation of memory. PLoS Pathog., 13, e1006544. https://doi.org/10.1371/journal.ppat.1006544. 34. Beaulieu, A., Zawislak, C., Nakayama, T. and Sun, J. (2014) The transcription factor Zbtb32 controls the proliferative burst of virus-specific natural killer cells responding to infection. Nat. Immunol., 15, 546–553. 35. Song, Y., Fukuda, N., Bai, C., Ma, T., Matthay, M. and Verkman, A. (2000) Role of aquaporins in alveolar fluid clearance in neonatal and adult lung, and in oedema formation following acute lung injury: studies in transgenic aquaporin null mice. J. Physiol., 525, 771–779. 36. Pires-Neto, R., Del Carlo Bernardi, F., Alves de Araujo, P., Mauad, T. and Dolhnikoff, M. (2016) The expression of water and ion channels in diffuse alveolar damage is not dependent on DAD Etiology. PLoS One, 11, e0166184. https://doi.org/10.1371/journal.pone.0166184. D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022 18 | Human Molecular Genetics, 2022, Vol. 00, No. 00 37. Ferrando, C., Suarez-Sipmann, F., Mellado-Artigas, R., Hernán- dez, M., Gea, A., Arruti, E., Aldecoa, C., Martínez-Pallí, G., Martínez-González, M., Slutsky, A. et al. (2020) Clinical features, ventilatory management, and outcome of ARDS caused by COVID-19 are similar to other causes of ARDS. Intensive Care Med., 46, 2200–2211. 38. Mariajoseph-Antony, L.F., Kannan, A., Panneerselvam, A., Loganathan, C., Anbarasu, K. and Prahalathan, C. (2020) Aquaporin water channels and lung physiology. Am. J. Physiol. Lung Cell Mol., 278, L867–L879. 39. Mariajoseph-Antony, L., Kannan, A., Panneerselvam, A., Loganathan, C., Anbarasu, K. and Prahalathan, C. (2020) Could aquaporin modulators be employed as prospective drugs for COVID-19 related pulmonary comorbidity? Med. Hypotheses, 143, 110201. https://doi.org/10.1016/j.mehy.2020.110201. 40. Charlesworth, B. (2001) Patterns of age-specific means and genetic variances of mortality rates predicted by the mutation- accumulation theory of ageing. J. Theor. Biol., 210, 47–65. 41. Perico, L., Benigni, A., Casiraghi, F., Ng, L., Renia, L. and Remuzzi, G. (2021) Immunity, endothelial injury and complement-induced coagulopathy in COVID-19. Nat. Rev. Nephrol., 17, 46–64. 42. Java, A., Apicelli, A., Liszewski, M., Coler-Reilly, A., Atkin- son, J., Kim, A. and Kulkarni, H. (2020) The complement sys- tem in COVID-19: friend and foe? JCI Insight, 5, e140711. https://doi.org/10.1172/jci.insight.140711. 43. Lo, M., Kemper, C. and Woodruff, T. (2020) COVID-19: com- plement, coagulation, and collateral damage. J. Inmunol., 205, 1488–1495. 44. Noris, M., Benigni, A. and Remuzzi, G. (2020) The case of comple- ment activation in COVID-19 multiorgan impact. Kidney Int., 98, 314–322. 45. Dunn, E. and Connor, J. (2012) Chapter 9—HijAkt: the PI3K/Akt pathway in virus replication and pathogenesis. Prog. Mol. Biol. Transl. Sci., 106, 223–250. 46. Lamy, P., Rébillard, X., Vacherot, F. and de la Taille, A. (2021) Androgenic hormones and the excess male mortality observed in COVID-19 patients: new convergent data. World J. Urol., 39, 3121–3123. 47. Wambier, C., Goren, A., Vaño-Galván, S., Müller, P., Ossimetha, A., Nau, G., Herrera, S. and McCoy, J. (2020) Androgen sensitivity gateway to COVID-19 disease severity.DrugDev.Res.,81, 771–776. 48. Shang,D.,Wang, L., Klionsky,D.,Cheng,H. and Zhou,R. (2021) Sex differences in autophagy-mediated diseases: toward precision medicine. Autophagy, 17, 1065–1076. 49. Wyce, A., Bai, Y., Nagpal, S. and Thompson, C. (2010) Research resource: the androgen receptor modulates expression of genes with critical roles in muscle development and function. Mol. Endocrinol., 24, 1665–1674. 50. Bettoun, D., Scafonas, A., Rutledge, S., Hodor, P., Chen, O., Gam- bone, C., Vogel, R., McElwee-Witmer, S., Bai, C., Freedman, L. et al. (2005) Interaction between the androgen receptor and RNase L mediates a cross-talk between the interferon and androgen signaling pathways. J. Biol. Chem., 280, 38898–38901. 51. Basrawala, Z., Alimirah, F., Xin, H., Mohideen, N., Campbell, S., Flanigan, R. and Choubey, D. (2006) Androgen receptor levels are increased by interferons in human prostate stromal and epithelial cells. Oncogene, 25, 2812–2817. 52. Lin, B., Ferguson, C., White, J., Wang, S., Vessella, R., True, L., Hood, L. and Nelson, P.S. (1999) Prostate-localized and androgen- regulated expression of the membrane-bound serine protease TMPRSS2. Cancer Res., 17, 4180–4184. 53. Huang, Q., Li, W., Zhou, Q., Awasthi, P., Cazin, C., Yap, Y., Mladenovic-Lucas, L., Hu, B., Jeyasuria, P., Zhang, L. et al. (2021) Leucine zipper transcription factor-like 1 (LZTFL1), an intraflagellar transporter protein 27 (IFT27) associated protein, is required for normal sperm function and male fertility. Dev. Biol., 477, 164–176. 54. Baldassarri, M., Picchiotti, N., Fava, F., Fallerini, C., Benetti, E., Daga, S., Valentino, F., Doddato, G., Furini, S., Gilib- erti, A. et al. (2021) Shorter androgen receptor polyQ alle- les protect against life-threatening COVID-19 disease in Euro- pean males. EBioMedicine, 65, 103246. https://doi.org/10.1016/j. ebiom.2021.103246. 55. Harris, P., Taylor, R., Thielke, R., Payne, J., Gonzalez, N. and Conde, J. (2009) Research electronic data capture (REDCap)—a metadata-drivenmethodology andworkflow process for provid- ing translational research informatics support. J. Biomed. Inform., 42, 377–381. 56. Harris, P., Taylor, R., Minor, B., Elliott, V., Fernandez, M., O’Neal, L., McLeod, L., Delacqua, G., Delacqua, F., Kirby, J. et al. (2019) The REDCap consortium: building an international com- munity of software partners. J. Biomed. Inform., 95, 103208. https://doi.org/10.1016/j.jbi.2019.103208. 57. The 1000 Genomes Project Consortium (2015) A global reference for human genetic variation. Nature, 526, 68–74. 58. Purcell, S., Ben, N., Todd-Brown, K., Thomas, L., Ferreira, M., Bender, D., Maller, J., Sklar, P., de Bakker, P., Daly, M. et al. (2007) PLINK: a tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet., 81, 559–575. 59. Graffelman, J. and Moreno, V. (2013) The mid P-value in exact tests for hardy-Weinberg equilibrium. Stat. Appl. Genet. Mol. Biol., 12, 433–448. 60. Price, A., Weale, M., Patterson, N., Myers, S., Need, A., Shianna, K., Ge, D., Rotter, J., Torres, E., Taylor, Q. et al. (2008) Long-range LD can confound genome scans in admixed populations. Am. J. Hum. Genet., 83, 132–139. 61. Alexander, D., Novembre, J. and Lange, K. (2009) Fast model- based estimation of ancestry in unrelated individuals. Genome Res., 19, 1655–1664. 62. Taliun, D., Harris, D., Kessler,M., Carlson, J., Szpiech, Z., Torres, R., Galiano Taliun, S., Corvelo, A., Gogarten, S., Kang, H. et al. (2021) Sequencing of 53,831 diverse genomes from the NHLBI TOPMed program. Nature, 590, 290–299. 63. Zheng, X., Levine, D., Shen, J., Gogarten, S., Laurie, C. and Weir, B. (2012) A high-performance computing toolset for relatedness and principal component analysis of SNP data. Bioinformatics,28, 3326–3328. 64. Zheng, X. and Davis,W. (2021) SAIGEgds—an efficient statistical tool for large-scale PheWAS with mixed models. Bioinformatics, 37, 728–730. 65. Zhou,W., Nielsen, J., Fritsche, L., Dey, R., Gabrielsen, M.,Wolford, B., LeFaive, J., VandeHaar, P., Gagliano, S., Figgord, A. et al. (2018) Efficiently controlling for case-control imbalance and sample relatedness in large-scale genetic association studies.Nat.Genet., 50, 1335–1341. 66. Willer, C., Li, Y. and Abecasis, G. (2010) METAL: fast and efficient meta-analysis of genomewide association scans. Bioinformatics, 26, 2190–2191. 67. Hutchinson, A., Watson, H. and Wallace, C. (2020) Improving the coverage of credible sets in Bayesian genetic fine-mapping. PLoS Comput. Biol., 16, e1007829. https://doi.org/10.1371/journal.pcbi.1007829. 68. Yang, J., Lee, H., Goddard, E. and Visscher, P. (2011) GCTA: a tool for genome-wide complex trait analysis. Am. J. Hum. Genet., 88, 76–82. 69. Templeton, A. and Read, B. (1994) Inbreeding: one word, several meanings, much confusion. EXS, 68, 91–105. D ow nloaded from https://academ ic.oup.com /hm g/advance-article/doi/10.1093/hm g/ddac132/6607933 by IN STITU TO D E SALU D C AR LO S III user on 19 O ctober 2022