The Tabloid Proteome: Orthogonal use of public proteomics data to derive biologically related protein network

Presented by Surya Gupta – Postdoctoral Researcher at VIB-UGent

Mass-spectrometry based proteomics experiments produces large amounts of data. While typically acquired to answer specific biological questions, these data can also be reused in orthogonal ways to reveal biological knowledge. We have developed a novel method for such orthogonal data reuse of public proteomics data to detect biologically associated protein pairs.

Mass-spectrometry proteomics experiments were obtained and reprocessed from the PRIDE database. For the identified proteins, we calculated the co-occurrence score, using Jaccard similarity. Protein pairs with score of atleast 0.4 were mapped to five knowledgebases; Reactome, Ensembl, IntAct, BioGRID, and CORUM, to assign potential biological relevance. Of the 2325 protein pairs that pass the Jaccard similarity threshold, we 81% of protein pairs with biological annotation (68% with five knowledgebases and 13% with GO terms). While comparison with randomly selected protein pairs, less than 2% protein pairs were found to be annotated. Furthermore, to extend the usability and accessibility of the detected protein pairs for research community, an online database called Tabloid Proteome was established.

Our approach shows that by re-using publically available data in a fully orthogonal way, effectively treating these data as a proteome-wide association study, we can extract various biologically meaningful patterns, which moreover, were quite complementary to associations detected by established protein-protein interaction techniques. Additionally, Tabloid Proteome features a simple yet powerful web interface that allows fast and easy access to all these protein associations, with their possible biological annotation.