Predictive interpretable models of cancer vulnerabilities via network based integration of multi-omic data
The development of new oncology drugs is at a critically high attrition level, and the number of molecular entities that are successfully developed into new therapeutic targets is constantly decreasing . This is most often due to a lack of robust genetic linkage between the trialled therapies and the disease they are designed for . To tackle this problem, large-scale pharmacological screenings have been performed recently across panels of immortalised human cancer cell lines [3, 4]. The drug-response datasets resulting from these studies have been integrated with multi-omic characterisations of the screened in vitro models, unveiling established and novel associations between cell molecular features and drug sensitivities [5–7]. The advent of genome editing by CRISPR-Cas9 has allowed extending these studies beyond the domain of currently druggable targets with precision and scale [8, 9]. Pooled CRISPR-Cas9 screenings employing genome-scale libraries of single guide RNAs are being performed on growing numbers of cancer in-vitro models [10–17]. This makes possible testing the extent to which inactivating each gene in turn impacts cancer viability in the context of a defined genomic/molecular make-up.
Datasets and results from these efforts are being increasingly released on the public domain, laying a rich preclinical foundation for the development of novel selective cancer targeted therapies.
By making use of these data, we will investigate how aggregating the multi-omic characterisation of the screened in vitro models might allow predicting their genetic-dependencies and responses to drug treatment. This will be achieved by leveraging deep-learning computational methods and prior-known large-scale pathway/signaling maps, and it will be oriented at producing mechanistically interpretable models, which might be translated into future therapeutic biomarkers and rules for potential combinatorial anti-cancer therapies.
Results from these efforts will be released through dedicated web-portals and beside software infrastructures allowing programmatic access. Most promising hits will be put forward for experimental validation with clinical/experimental collaborators.
2. Cook D, Brown D, Alexander R, March R, Morgan P, Satterthwaite G, et al. Lessons learned from the fate of AstraZeneca’s drug pipeline: a five-dimensional framework. Nat Rev Drug Discov. 2014;13:419–31.
3. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature. 2012;483:603–7.
6. Basu A, Bodycombe NE, Cheah JH, Price EV, Liu K, Schaefer GI, et al. An interactive resource to identify cancer genetic and lineage dependencies targeted by small molecules. Cell. 2013;154:1151–61.
8. Evers B, Jastrzebski K, Heijmans JPM, Grernrum W, Beijersbergen RL, Bernards R. CRISPR knockout screening outperforms shRNA and CRISPRi in identifying essential genes. Nat Biotechnol. 2016;34:631–3.
12. Tzelepis K, Koike-Yusa H, De Braekeleer E, Li Y, Metzakopian E, Dovey OM, et al. A CRISPR Dropout Screen Identifies Genetic Vulnerabilities and Therapeutic Targets in Acute Myeloid Leukemia. Cell Rep. 2016;17:1193–205.
17. Meyers RM, Bryan JG, McFarland JM, Weir BA. Computational correction of copy number effect improves specificity of CRISPR–Cas9 essentiality screens in cancer cells. Nature. 2017. https://www.nature.com/articles/ng.3984.