Network Biology-Inspired Machine Learning Features Predict Cancer Gene Targets and Reveal Target Coordinating Mechanisms

Abstract

Anticipating and understanding cancers’ need for specific gene activities is key for novel therapeutic development. Here we utilized DepMap, a cancer gene dependency screen, to demonstrate that machine learning combined with network biology can produce robust algorithms that both predict what genes a cancer is dependent on and what network features coordinate such gene dependencies. Using network topology and biological annotations, we constructed four groups of novel engineered machine learning features that produced high accuracies when predicting binary gene dependencies. We found that in all examined cancer types, F1 scores were greater than 0.90, and model accuracy remained robust under multiple hyperparameter tests. We then deconstructed these models to identify tumor type-specific coordinators of gene dependency and identified that in certain cancers, such as thyroid and kidney, tumors’ dependencies are highly predicted by gene connectivity. In contrast, other histologies relied on pathway-based features such as lung, where gene dependencies were highly predictive by associations with cell death pathway genes. In sum, we show that biologically informed network features can be a valuable and robust addition to predictive pharmacology models while simultaneously providing mechanistic insights.

Modeling cancer dependency with deep graph models

“…Mining the relationship between gene expression, CNV, mutations, etc., and cancer gene dependencies through machine learning can not only help reveal the mechanism of cancer gene dependencies, but more importantly, also provide an efficient method to establish personalized cancer-dependent gene maps from tumor patient omics data, which is of great value for promoting precision cancer diagnosis and treatment. Currently, majority of the available methods still rely on traditional machine learning methods, and could not generalize well to unseen samples [18][19][20][21] 。 Currently only one deep learning method based on autoencoders, deepDep, has been developed to predict cancer dependencies 22 . Although deepDep achieved good prediction on the BROAD CRISPR dataset (test set correlation coefficient 0.87).…”