|
Description: This page provides the data and predictions utilized and generated in our study: Yin R, Pierce BG. Evaluation of AlphaFold Antibody-Antigen Modeling with Implications for Improving Predictive Accuracy. bioRxiv 2023.07.05.547832 bioRxiv preprint Data download: Curated PDB structures of antibody-antigen cases used for benchmarking (download structures). The chains in structures have been renamed and organized in the following sequence: antibody heavy and light chains (if present) followed by antigen chains. The antibody was trimmed to the variable domain, and renumbered the AHo numbering scheme. AlphaFold v.2.2.0 predictions of 429 antibody-antigen complexes (download predictions), and the MSA files of these predictions (download MSAs). ColabFold predictions of 428 antibody-antigen complexes. Since the files are large, we offer to download them in parts: part 1, part 2, part 3, part 4, part 5, and scores. AlphaFold v.2.2.0 predictions of a subset of 100 antibody-antigen complexes, using different templates (download predictions). AlphaFold v.2.3.0 predictions of 39 antibody-antigen complexes (download predictions). Inside the tarball, you will find AlphaFold v2.3.0 predictions generated with default settings, and reduced maximum recycling interations (Ncycle), and AlphaFold (v.2.2.0) predictions of the same cases, using a date cutoff matching the setting used in generating v.2.3.0 predictions. |