Fig. 1: Overview of CF-random. | Nature Communications

Fig. 1: Overview of CF-random.

From: Large-scale predictions of alternative protein conformations by AlphaFold2-based sequence association

Fig. 1

CF-random (ColabFold-based random MSA sampling) generates dominant and alternative protein structures by combining ColabFold predictions generated from a deep multiple sequence alignment (MSA) and shallow random MSAs, respectively. To run default mode, a full-length MSA is required. ColabFold samples the MSA at default depth (512:5120) to produce the dominant conformation and then randomly subsamples at shallower depths to predict putative alternative conformations. It was benchmarked against known fold-switching proteins by calculating the TM-scores of all predicted structures against two reference structures. Success was considered correct predictions of both conformations from a single target sequence. Source data are provided as a Source Data file.

Back to article page