LOTUS: a Single- and Multitask Machine Learning Algorithm for the Prediction of Cancer Driver Genes

Abstract

Cancer driver genes, i.e., oncogenes and tumor suppressor genes, are involved in the acquisition of important functions in tumors, providing a selective growth advantage, allowing uncontrolled proliferation and avoiding apoptosis. It is therefore important to identify these driver genes, both for the fundamental understanding of cancer and to help finding new therapeutic targets. Although the most frequently mutated driver genes have been identified, it is believed that many more remain to be discovered, particularly for driver genes specific to some cancer types.In this paper we propose a new computational method called LOTUS to predict new driver genes. LOTUS is a machine-learning based approach which allows to integrate various types of data in a versatile manner, including informations about gene mutations and protein-protein interactions. In addition, LOTUS can predict cancer 1 driver genes in a pan-cancer setting as well as for specific cancer types, using a multitask learning strategy to share information across cancer types.We empirically show that LOTUS outperforms three other state-of-the-art driver gene prediction methods, both in terms of intrinsic consistency and prediction accuracy, and provide predictions of new cancer genes across many cancer types. Author summaryCancer development is driven by mutations and dysfunction of important, so-called driver genes in biological networks. We propose LOTUS, a new computational approach 7 to identify genes with high oncogenic potential. LOTUS implements a machine learning 8 approach to learn an oncogenic potential score from known driver genes, and brings two 9 novelties compared to existing methods. First, it allows to easily combine heterogeneous 10 informations into the scoring function, which we illustrate by learning a scoring function 11 from both known mutations in large cancer cohorts and interactions in biological 12 networks. Second, using a multitask learning strategy, it can predict different driver 13 genes for different cancer types, while sharing information between them to improve the 14 prediction for every type. We provide experimental results showing that LOTUS 15 significantly outperforms several state-of-the-art cancer gene prediction softwares.

AeQTL: eQTL analysis using region-based aggregation of rare genomic variants

“…Cancer driver genes, depending on their mutated cancer type and pathway context, can be subclassified into oncogenes and tumor suppressor genes (TSGs). But most existing methods to classify oncogenes and TSGs leveraged cohort-level mutation data 22 – 25 that lack considerations of their downstream consequences. To understand whether eQTL patterns could capture the distinction between oncogenes and TSGs, we further investigated the significant genes classified as oncogene or TSG from Bailey et al’s DNA mutation-based study 16 .…”

“…Traditional methods to classify oncogenes or tumor suppressors rely on algorithms considering only DNA-mutation patterns or functional curation 22 – 25 . Herein, we present truncation eQTL patterns revealed by AeQTL as a potential new method to distinguish oncogenes (elevated expression) from tumor suppressors (reduced expression).…”

Artificial intelligence and machine learning‐aided drug discovery in central nervous system diseases: State‐of‐the‐arts and future directions

“…Multitask learning has also been employed for the prediction of cancer driver genes. LOTUS, an ML‐based algorithm, predicts cancer driver genes in a pan‐cancer setting, as well as for specific cancer types, using a multitask learning strategy sharing information across cancer types 48 . For the readers who want to learn more about opportunities and challenges in predictive modeling for multiomics data sets, we suggest the review paper of Kim and Tagkopoulos 49 …”

Integrative Data Analytic Framework to Enhance Cancer Precision Medicine

“…We compare the performance of our method with the following state-of-the-art methods: NMFR, network-based integration (NBI), 32 LOTUS, 33 and Subdyquency. 34 The methods were developed to predict cancer-related genes in slightly different contexts and are adapted to our problem here (see Supplementary Methods section for details).…”