Functionathon: a manual data mining workflow to generate functional hypotheses for uncharacterized human proteins and its application by undergraduate students.

Paula Duek,Amos Bairoch,Camille Mary,Lydie Lane,Monique Zahn-Zabal

doi:10.1093/database/baab046

Abstract

About 10% of human proteins have no annotated function in protein knowledge bases. A workflow to generate hypotheses for the function of these uncharacterized proteins has been developed, based on predicted and experimental information on protein properties, interactions, tissular expression, subcellular localization, conservation in other organisms, as well as phenotypic data in mutant model organisms. This workflow has been applied to seven uncharacterized human proteins (C6orf118, C7orf25, CXorf58, RSRP1, SMLR1, TMEM53 and TMEM232) in the frame of a course-based undergraduate research experience named Functionathon organized at the University of Geneva to teach undergraduate students how to use biological databases and bioinformatics tools and interpret the results. C6orf118, CXorf58 and TMEM232 were proposed to be involved in cilia-related functions; TMEM53 and SMLR1 were proposed to be involved in lipid metabolism and C7orf25 and RSRP1 were proposed to be involved in RNA metabolism and gene expression. Experimental strategies to test these hypotheses were also discussed. The results of this manual data mining study may contribute to the project recently launched by the Human Proteome Organization (HUPO) Human Proteome Project aiming to fill gaps in the functional annotation of human proteins.Database URL: http://www.nextprot.org

Highlights

Biomedical research is an area generating a prodigious amount of information obtained from massive datasets
A workflow to generate hypotheses for the function of these uncharacterized proteins has been developed, based on predicted and experimental information on protein properties, interactions, tissular expression, subcellular localization, conservation in other organisms, as well as phenotypic data in mutant model organisms. This workflow has been applied to seven uncharacterized human proteins (C6orf118, C7orf25, CXorf58, RSRP1, SMLR1, TMEM53 and TMEM232) in the frame of a course-based undergraduate research experience named Functionathon organized at the University of Geneva to teach undergraduate students how to use biological databases and bioinformatics tools and interpret the results
C6orf118, CXorf58 and TMEM232 were proposed to be involved in cilia-related functions; TMEM53 and SMLR1 were proposed to be involved in lipid metabolism and C7orf25 and RSRP1 were proposed to be involved in RNA metabolism and gene expression

Summary

Introduction

Biomedical research is an area generating a prodigious amount of information obtained from massive datasets. Databases organizing, standardizing and distributing this information, as well as bioinformatics tools allowing to mine these data, so as to translate it into knowledge useful for discovery are crucial to modern biological research. We have shown that it was possible to propose hypotheses for the function of such proteins by combining information from the literature, experimental repositories, databases and bioinformatics tools [3,4,5]. These proposed functions have been published and are awaiting experimental validation to be annotated in neXtProt

Methods

Results

Conclusion