CEN‐tools: an integrative platform to identify the contexts of essential genes

Abstract

An emerging theme from large‐scale genetic screens that identify genes essential for cell fitness is that essentiality of a given gene is highly context‐specific. Identification of such contexts could be the key to defining gene function and also to develop novel therapeutic interventions. Here, we present Context‐specific Essentiality Network‐tools (CEN‐tools), a website and python package, in which users can interrogate the essentiality of a gene from large‐scale genome‐scale CRISPR screens in a number of biological contexts including tissue of origin, mutation profiles, expression levels and drug responses. We show that CEN‐tools is suitable for the systematic identification of genetic dependencies and for more targeted queries. The associations between genes and a given context are represented as dependency networks (CENs), and we demonstrate the utility of these networks in elucidating novel gene functions. In addition, we integrate the dependency networks with existing protein–protein interaction networks to reveal context‐dependent essential cellular pathways in cancer cells. Together, we demonstrate the applicability of CEN‐tools in aiding the current efforts to define the human cellular dependency map.

CoRe: A robustly benchmarked R package for identifying core-fitness genes in genome-wide pooled CRISPR-Cas9 screens

“…We compared the sets of CFGs and CEGs predicted by CoRe when applied to the largest integrative dataset of cancer dependency assembled to date [20] with state-of-the-art sets of core-fitness genes derived from recent functional genetic screening datasets [10,12,18,19], as well as with the output of a logistic-regression based method, part of the recent CEN-tools software proposed in [18], applied to the same dataset [20].…”

“…We downloaded the CEN-tools package [18] from https://gitlab.ebi.ac.uk/petsalakilab/centools/-/tree/master/CEN-tools. In order to decrease the memory burden for the GitHub repository of the CoRe package, we removed all the python modules and data objects that were not directly called by the LR.py and clustering.R functions, respectively the python script implementing the logistic regression model and the R script performing the subsequent cluster analysis.…”

“…This method builds on the basic intuition that if a gene is universally essential then it should rank among the top essential genes in the vast majority of screened models, including those that are the least dependant on it, or generally showing a moderate to weak loss-of-fitness phenotype upon CRISPR-Cas9 targeting. Finally, a logistic regression based method for classifying genes into CFGs or context specific essentials has been recently introduced by Sharma and colleagues [18] as part of the CEN-tools suite, using reference sets of essential and non-essential genes for the training phase [19].…”

“…We present CoRe : an R package implementing recently proposed as well as novel versions of algorithms for the identification of CFGs from a joint analysis of multiple genome-wide pooled CRISPR-Cas9 knock-out screens. Furthermore, we present results from a comparison of CoRe’s output (when applied to the largest integrative cancer dependency dataset generated to date [20]) against widely used [10,19], or more recent [18] sets of CFGs obtained via an alternative approach (also tested on the same recent cancer dependency dataset). We report an increased coverage of prior known human essential genes, new potential core-fitness genes, and lower false positive rates for CoRe’s methods with respect to other state-of-the-art core-fitness sets and available methods.…”

Dynamic rewiring of biological activity across genotype and lineage revealed by context-dependent functional interactions

“…Many such modules found in the network encode well annotated biological processes, enabling robust functional prediction for uncharacterized genes (Aregger et al, 2020; Noordermeer et al, 2018; Zimmermann et al, 2018). Some modules, however, are associated with specific tissues or genomic perturbations, such as the ESR1-FOXA1 pathway in ER+ breast cancer cell lines and the BRAF V600E module in melanoma cells (Amici et al, 2020; Kim et al, 2019; Sharma et al, 2020). In this latter cluster, gene mutation is so conflated with tissue specificity that b-RAF is tightly linked with melanocyte-specific transcription factors MITF and SOX9 but does not show significant correlation with its known protein interaction partner c-RAF (encoded by the RAF1 gene).…”