Sewon Lee1,10, Gyuri Kim1,10, Eli Levy Karin2, Milot Mirdita1, Sukhwan Park3, Rayan Chikhi4, Artem Babaian5,6, Andriy Kryshtafovych7 and Martin Steinegger1,3,8,9 1School of Biological Sciences, Seoul National University, Gwanak-gu, Seoul 08826, South Korea 2ELKMO, Copenhagen 2720, Denmark 3Interdisciplinary Program in Bioinformatics, Seoul National University, Seoul 08826, South Korea 4Institut Pasteur, Université Paris Cité, G5 Sequence Bioinformatics, 75015 Paris, France 5Department of Molecular Genetics, University of Toronto, Toronto, Ontario M5S 1A8, Canada 6Donnelly Centre for Cellular and Biomolecular Research, University of Toronto, Toronto, Ontario M5S 3E1, Canada 7Genome Center, University of California, Davis, California 95616, USA 8Artificial Intelligence Institute, Seoul National University, Seoul 08826, South Korea 9Institute of Molecular Biology and Genetics, Seoul National University, Seoul 08826, South Korea Correspondence: martin.steineggersnu.ac.kr

↵10 These authors contributed equally to this work.

The recent CASP15 competition highlighted the critical role of multiple sequence alignments (MSAs) in protein structure prediction, as demonstrated by the success of the top AlphaFold2-based prediction methods. To push the boundaries of MSA utilization, we conducted a petabase-scale search of the Sequence Read Archive (SRA), resulting in gigabytes of aligned homologs for CASP15 targets. These were merged with default MSAs produced by ColabFold-search and provided to ColabFold-predict. By using SRA data, we achieved highly accurate predictions (GDT_TS > 70) for 66% of the non-easy targets, whereas using ColabFold-search default MSAs scored highly in only 52%. Next, we tested the effect of deep homology search and ColabFold's advanced features, such as more recycles, on prediction accuracy. While SRA homologs were most significant for improving ColabFold's CASP15 ranking from 11th to 3rd place, other strategies contributed too. We analyze these in the context of existing strategies to improve prediction.