ResponseNet2.0: Revealing signaling and regulatory pathways connecting your proteins and genes--now with human data

Abstract

Genome sequencing and transcriptomic profiling are two widely used approaches for the identification of human disease pathways. However, each approach typically provides a limited view of disease pathways: Genome sequencing can identify disease-related mutations but rarely reveals their mode-of-action, while transcriptomic assays do not reveal the series of events that lead to the transcriptomic change. ResponseNet is an integrative network-optimization approach that we developed to fill these gaps by highlighting major signaling and regulatory molecular interaction paths that connect disease-related mutations and genes. The ResponseNet web-server provides a user-friendly interface to ResponseNet. Specifically, users can upload weighted lists of proteins and genes and obtain a sparse, weighted, molecular interaction subnetwork connecting them, that is biased toward regulatory and signaling pathways. ResponseNet2.0 enhances the functionality of the ResponseNet web-server in two important ways. First, it supports analysis of human data by offering a human interactome composed of proteins, genes and micro-RNAs. Second, it offers a new informative view of the output, including a randomization analysis, to help users assess the biological relevance of the output subnetwork. ResponseNet2.0 is available at http://netbio.bgu.ac.il/respnet .

Figure 1.

ResponseNet sensitivity and precision upon analysis of known human pathways. Results are shown for full maps (denoted ‘Full’) and for different path lengths (lengths >12 involved <5 maps and were therefore ignored). Each box-plot diagram shows the quartile values (25, 50 and 75%). (A) Nodes sensitivity. (B) Nodes precision. (C) Interactions sensitivity. (D) Interactions precision. Specificity was 99% in every case and is therefore not shown.

Figure 2.

An example of ResponseNet2.0 output. The human protein BAP1 is a deubiquitinating enzyme that acts as a regulator of cell growth by mediating deubiquitination of HCFC1, and is frequently mutated in uveal melanomas. We used ResponseNet2.0 to identify a regulatory subnetwork that connects BAP1 to human genes that were found to be down-regulated in melanoma cell lines. ResponseNet correctly predicted the BAP1-interacting protein HCFC1, and connected HCFC1 to two transcription factors, GABPA and SP1, that regulate the transcription of 16 target genes. Notably, SP1 was previously linked to melanoma (33). S and T are auxiliary nodes that are part of ResponseNet formulation.