A Novel System for Functional Determination of Variants of Uncertain Significance using Deep Convolutional Neural Networks

Lior Zimmerman,Michael Vidne,Zohar Barbash,Arie Aizenmann,Ori Zelichov,Gabi Tarcic

doi:10.1038/s41598-020-61173-1

Abstract

Many drugs are developed for commonly occurring, well studied cancer drivers such as vemurafenib for BRAF V600E and erlotinib for EGFR exon 19 mutations. However, most tumors also harbor mutations which have an uncertain role in disease formation, commonly called Variants of Uncertain Significance (VUS), which are not studied or characterized and could play a significant role in drug resistance and relapse. Therefore, the determination of the functional significance of VUS and their response to Molecularly Targeted Agents (MTA) is essential for developing new drugs and predicting response of patients. Here we present a multi-scale deep convolutional neural network (DCNN) architecture combined with an in-vitro functional assay to investigate the functional role of VUS and their response to MTA’s. Our method achieved high accuracy and precision on a hold-out set of examples (0.98 mean AUC for all tested genes) and was used to predict the oncogenicity of 195 VUS in 6 genes. 63 (32%) of the assayed VUS’s were classified as pathway activating, many of them to a similar extent as known driver mutations. Finally, we show that responses of various mutations to FDA approved MTAs are accurately predicted by our platform in a dose dependent manner. Taken together this novel system can uncover the treatable mutational landscape of a drug and be a useful tool in drug development.

Highlights

Given the abundance of Variants of Uncertain Significance (VUS) in these datasets, a strategy that includes accurate characterization of the activity of VUS and their response to Molecularly Targeted Agents (MTA) could provide significant benefit to drug development and increase the success rates of clinical trials
Out of the data set of 7 genes, for which we have a total of 301 mutated variants and wildtype forms of each we selected 8 mutated variants to be used as positive examples of pathway activation and the wildtype form as negative examples for pathway activation
We tested whether there was a difference in the pathway activation patterns induced by each gene by training the deep convolutional neural network (DCNN) to predict with which gene the cells in the image were transfected

Summary

Introduction

Given the abundance of VUS in these datasets, a strategy that includes accurate characterization of the activity of VUS and their response to MTAs could provide significant benefit to drug development and increase the success rates of clinical trials. The performance of the algorithm is high (AUC of 0.984 and 0.976 on breast cancer and colorectal cancer), predicting the role of individual mutations and VUS in particular, is beyond its scope Another tool, Mut2Vec[27], is an unsupervised approach for cancer driver prediction which is based on the popular Word2Vec[28] class of models. In Mut2Vec, the model is trained on a set of cancer profiles to generate an embedding for each mutation, showing that passenger and driver mutations can be distinguished when the embeddings are clustered Pathology is another field of cancer research that has gone through significant transformations by the recent advances in deep learning[29].

Methods

Results

Conclusion