Herson H M Soares, João P R Romanelli, Patrick J Fleming, Carlos H da Silveira
{"title":"FIBOS: R and python packages for analyzing protein packing and structure.","authors":"Herson H M Soares, João P R Romanelli, Patrick J Fleming, Carlos H da Silveira","doi":"10.1093/bioinformatics/btaf434","DOIUrl":null,"url":null,"abstract":"<p><strong>Motivation: </strong>Advances in the prediction of the 3D structures of most known proteins through machine learning have achieved unprecedented accuracies. However, although these computed models are remarkably good, they still challenge accuracy at the atomic level. The Occluded Surface (OS) algorithm is widely used for atomic packing analysis. But it lacks implementations in high-level languages.</p><p><strong>Results: </strong>We introduce FIBOS, an R and Python package incorporating the OS methodology with enhancements. We show how FIBOS can be used to atomically compare experimental structures and AlphaFold predictions. Although the average packing was similar, AlphaFold models exhibited slightly greater variability, revealing a specific pattern of outliers.</p><p><strong>Availability and implementation: </strong>FIBOS can be installed locally as a PyPi Python or CRAN R package, and it is also available at https://github.com/insilico-unifei/fibos-R and https://github.com/insilico-unifei/fibos-py.</p>","PeriodicalId":93899,"journal":{"name":"Bioinformatics (Oxford, England)","volume":" ","pages":""},"PeriodicalIF":5.4000,"publicationDate":"2025-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC12449057/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Bioinformatics (Oxford, England)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1093/bioinformatics/btaf434","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Motivation: Advances in the prediction of the 3D structures of most known proteins through machine learning have achieved unprecedented accuracies. However, although these computed models are remarkably good, they still challenge accuracy at the atomic level. The Occluded Surface (OS) algorithm is widely used for atomic packing analysis. But it lacks implementations in high-level languages.
Results: We introduce FIBOS, an R and Python package incorporating the OS methodology with enhancements. We show how FIBOS can be used to atomically compare experimental structures and AlphaFold predictions. Although the average packing was similar, AlphaFold models exhibited slightly greater variability, revealing a specific pattern of outliers.
Availability and implementation: FIBOS can be installed locally as a PyPi Python or CRAN R package, and it is also available at https://github.com/insilico-unifei/fibos-R and https://github.com/insilico-unifei/fibos-py.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
