Ontology based molecular signatures for immune cell types via gene expression analysis
- Equal contributors
1 European Molecular Biology Laboratory, European Bioinformatics Institute, Wellcome Trust Genome Campus, Hinxton, Cambridge CB10 1SD, UK
2 Ontology Development Group, Library, Oregon Health and Science University, 3181 SW Sam Jackson Park Road, Portland, Oregon 97239, USA
3 Genomics Division, Lawrence Berkeley National Laboratory, 1 Cyclotron Road, Berkeley, CA 94720, USA
4 Baylor Institute for Immunology Research, Dallas, TX 75204, USA
5 Bioinformatics and Computational Biology, The Jackson Laboratory, 600 Main Street, Bar Harbor, ME 04609, USA
6 Department of Neurology, University at Buffalo School of Medicine and Biomedical Sciences, 100 High Street, Buffalo, NY 14203, USA
BMC Bioinformatics 2013, 14:263 doi:10.1186/1471-2105-14-263Published: 30 August 2013
New technologies are focusing on characterizing cell types to better understand their heterogeneity. With large volumes of cellular data being generated, innovative methods are needed to structure the resulting data analyses. Here, we describe an ‘Ontologically BAsed Molecular Signature’ (OBAMS) method that identifies novel cellular biomarkers and infers biological functions as characteristics of particular cell types. This method finds molecular signatures for immune cell types based on mapping biological samples to the Cell Ontology (CL) and navigating the space of all possible pairwise comparisons between cell types to find genes whose expression is core to a particular cell type’s identity.
We illustrate this ontological approach by evaluating expression data available from the Immunological Genome project (IGP) to identify unique biomarkers of mature B cell subtypes. We find that using OBAMS, candidate biomarkers can be identified at every strata of cellular identity from broad classifications to very granular. Furthermore, we show that Gene Ontology can be used to cluster cell types by shared biological processes in order to find candidate genes responsible for somatic hypermutation in germinal center B cells. Moreover, through in silico experiments based on this approach, we have identified genes sets that represent genes overexpressed in germinal center B cells and identify genes uniquely expressed in these B cells compared to other B cell types.
This work demonstrates the utility of incorporating structured ontological knowledge into biological data analysis – providing a new method for defining novel biomarkers and providing an opportunity for new biological insights.