Pathway-specific model estimation for improved pathway annotation by network crosstalk

Abstract

Pathway enrichment analysis is the most common approach for understanding which biological processes are affected by altered gene activities under specific conditions. However, it has been challenging to find a method that efficiently avoids false positives while keeping a high sensitivity. We here present a new network-based method ANUBIX based on sampling random gene sets against intact pathway. Benchmarking shows that ANUBIX is considerably more accurate than previous network crosstalk based methods, which have the drawback of modelling pathways as random gene sets. We demonstrate that ANUBIX does not have a bias for finding certain pathways, which previous methods do, and show that ANUBIX finds biologically relevant pathways that are missed by other methods.

Network Crosstalk as a Basis for Drug Repurposing

“…Except for the ranking on the Guney2016 benchmark, it was clear that the top crosstalk-based approach was ANUBIX . This can be attributed to the fact that the both BinoX and NEAT approaches are more prone to false positives compared with ANUBIX due to the beta-binomial distribution being a more accurate model of random crosstalk in the underlying network and that ANUBIX accounts for nonrandom intrapathway interactions ( Castresana-Aguirre and Sonnhammer, 2020 ). An additional insight from the three benchmarks was that all methods performed much better on the FCbench, which can be attributed to the more comprehensive underlying functional association network, FunCoup.…”

“…The mean and standard deviation of the sampled crosstalk are calculated as for BinoX and a similar normalization procedure is applied to obtain the normalized crosstalk score, . The estimation of parameters for the Beta distribution as well as the calculation of mean, standard deviation, and p -values was conducted using a re-implementation of the pathway enrichment tool ANUBIX ( Castresana-Aguirre and Sonnhammer, 2020 ).…”

“…The newest methods, in pathway analysis, study network crosstalk, i.e., the interaction between sets of genes in the network, for example, the crosstalk between a set of disease-related genes and pathways with known biological functions. Given the substantial performance increase of crosstalk-based pathway analysis compared with other approaches ( Ogris et al, 2017 ; Castresana-Aguirre and Sonnhammer, 2020 ), we propose to use crosstalk-based tools also for drug repurposing. One way to achieve this is to evaluate the similarity between drug-target proteins and disease-related proteins by measuring the crosstalk between the two sets compared with an expected crosstalk in the underlying network between similar sets of nodes in terms of size and connectivity.…”

ViralLink: An integrated workflow to investigate the effect of SARS-CoV-2 on intracellular signalling and regulatory pathways

“…Furthermore, ViralLink also implements a network-based pathway enrichment analysis on the upstream human proteins of the network using the ANUBIX (Adaptive NUll distriButIon of X-talk) algorithm [ 37 ]. Here, limitations of overrepresentation analysis, such as the assumption of gene independency, are overcome by integrating knowledge of gene/protein associations using networks.…”

“…Finally, the ViralLink workflow employs functional analysis and visualisation methods to aid interpretation of the generated intracellular networks, enabling detailed investigation of key proteins and signalling pathways. Regarding functional analysis, both overrepresentation analysis and a network-aware pathway algorithm called ANUBIX are employed, considering Reactome, Gene Ontology and KEGG annotations [37,[55][56][57]. This varied approach is taken to avoid biases due to variability in results output using different functional analysis tools and functional databases [58].…”