ColabFold output files for the paper: "Flexibility or uncertainty? A critical assessment of AlphaFold 2 pLDDT"

Dataset description This repository contains the ColabFold (AlphaFold 2) output data used in the article “Flexibility or uncertainty? A critical assessment of AlphaFold 2 pLDDT”, published in Structure (Cell Press). These files enable full reproducibility of the structural predictions and pLDDT values, and include all multiple sequence alignments generated during execution. The protein set is based on the first version of the ATLAS dataset (Database: https://www.dsimb.inserm.fr/ATLAS, Article: https://doi.org/10.1093/nar/gkad1084). For each protein, the folder includes: MSA files: *.a3m (input MSAs, including bfd.mgnify30.metaeuk30.smag30.a3m, uniref.a3m) pdb70.m8 (template search results) out.tar.gz (compressed MSA data) Configuration and logs: config.json (run parameters) log.txt (execution log) cite.bibtex (software citation) Prediction outputs: *_relaxed_rank_*.pdb and *_unrelaxed_rank_*.pdb (predicted structures) *_scores_rank_*.json (model scores and metrics) *_predicted_aligned_error_v1.json (PAE values) *_template_domain_names.json Visual summaries: *_plddt.png (pLDDT per residue) *_pae.png (Predicted Aligned Error matrix) *_coverage.png (MSA coverage) Note: PDB identifiers for all structures are included in the provided file names. .cif files were omitted due to file size limitations, but can be regenerated from the PDB IDs in *_template_domain_names.json if needed.