Alexander Kroll, Nico Niebuhr, Gregory Butler, Martin J Lercher
{"title":"SPOT: A machine learning model that predicts specific substrates for transport proteins.","authors":"Alexander Kroll, Nico Niebuhr, Gregory Butler, Martin J Lercher","doi":"10.1371/journal.pbio.3002807","DOIUrl":null,"url":null,"abstract":"<p><p>Transport proteins play a crucial role in cellular metabolism and are central to many aspects of molecular biology and medicine. Determining the function of transport proteins experimentally is challenging, as they become unstable when isolated from cell membranes. Machine learning-based predictions could provide an efficient alternative. However, existing methods are limited to predicting a small number of specific substrates or broad transporter classes. These limitations stem partly from using small data sets for model training and a choice of input features that lack sufficient information about the prediction problem. Here, we present SPOT, the first general machine learning model that can successfully predict specific substrates for arbitrary transport proteins, achieving an accuracy above 92% on independent and diverse test data covering widely different transporters and a broad range of metabolites. SPOT uses Transformer Networks to represent transporters and substrates numerically. To overcome the problem of missing negative data for training, it augments a large data set of known transporter-substrate pairs with carefully sampled random molecules as non-substrates. SPOT not only predicts specific transporter-substrate pairs, but also outperforms previously published models designed to predict broad substrate classes for individual transport proteins. We provide a web server and Python function that allows users to explore the substrate scope of arbitrary transporters.</p>","PeriodicalId":49001,"journal":{"name":"PLoS Biology","volume":null,"pages":null},"PeriodicalIF":9.8000,"publicationDate":"2024-09-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11426516/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"PLoS Biology","FirstCategoryId":"99","ListUrlMain":"https://doi.org/10.1371/journal.pbio.3002807","RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2024/9/1 0:00:00","PubModel":"eCollection","JCR":"Q1","JCRName":"Agricultural and Biological Sciences","Score":null,"Total":0}
引用次数: 0
Abstract
Transport proteins play a crucial role in cellular metabolism and are central to many aspects of molecular biology and medicine. Determining the function of transport proteins experimentally is challenging, as they become unstable when isolated from cell membranes. Machine learning-based predictions could provide an efficient alternative. However, existing methods are limited to predicting a small number of specific substrates or broad transporter classes. These limitations stem partly from using small data sets for model training and a choice of input features that lack sufficient information about the prediction problem. Here, we present SPOT, the first general machine learning model that can successfully predict specific substrates for arbitrary transport proteins, achieving an accuracy above 92% on independent and diverse test data covering widely different transporters and a broad range of metabolites. SPOT uses Transformer Networks to represent transporters and substrates numerically. To overcome the problem of missing negative data for training, it augments a large data set of known transporter-substrate pairs with carefully sampled random molecules as non-substrates. SPOT not only predicts specific transporter-substrate pairs, but also outperforms previously published models designed to predict broad substrate classes for individual transport proteins. We provide a web server and Python function that allows users to explore the substrate scope of arbitrary transporters.
期刊介绍:
PLOS Biology is the flagship journal of the Public Library of Science (PLOS) and focuses on publishing groundbreaking and relevant research in all areas of biological science. The journal features works at various scales, ranging from molecules to ecosystems, and also encourages interdisciplinary studies. PLOS Biology publishes articles that demonstrate exceptional significance, originality, and relevance, with a high standard of scientific rigor in methodology, reporting, and conclusions.
The journal aims to advance science and serve the research community by transforming research communication to align with the research process. It offers evolving article types and policies that empower authors to share the complete story behind their scientific findings with a diverse global audience of researchers, educators, policymakers, patient advocacy groups, and the general public.
PLOS Biology, along with other PLOS journals, is widely indexed by major services such as Crossref, Dimensions, DOAJ, Google Scholar, PubMed, PubMed Central, Scopus, and Web of Science. Additionally, PLOS Biology is indexed by various other services including AGRICOLA, Biological Abstracts, BIOSYS Previews, CABI CAB Abstracts, CABI Global Health, CAPES, CAS, CNKI, Embase, Journal Guide, MEDLINE, and Zoological Record, ensuring that the research content is easily accessible and discoverable by a wide range of audiences.