题目： FAS: Assessing the similarity between proteins using multi-layered feature architectures
Document source: https://doi.org/10.1101/2022.09.01.506207 (biorxiv)
Code: https://pypi.org/project/greedyFAS/
Introduction: Protein sequence comparison is an essential element in the bioinformatics toolkit. When sequences are annotated with features such as functional domains, transmembrane domains, regions of low complexity, or secondary structure elements, the resulting characteristic structures allow better informed comparisons. However, many existing architectural similarity scoring schemes fail to handle features produced by multiple annotation sources, especially those that address overlapping and redundant feature annotations. This paper presents FAS, a scoring method that integrates features from multiple annotation sources into a directed acyclic architecture graph. Find a way to solve the redundancy of the architecture comparison part by realizing the graph that maximizes the similarity of the pairwise architectures. In large-scale assessments of more than 10,000 human-yeast homology pairs, FAS-assessed structural similarities were consistently more confident than overlaps resolved or unresolved using e-values. Three case studies demonstrate the utility of FAS on structural comparison tasks: benchmarking of homolog assignment software, identification of functionally differentiated homologues, and diagnosis of protein structural changes caused by erroneous gene predictions. With the help of FAS, feature schema comparisons can now be routinely integrated into these and many other applications.
main content:
-------------------------------------------
Welcome to like, collect and forward!
See you next time!