NeuMF: Predicting Anti-cancer Drug Response Through a Neural Matrix Factorization Model

Abstract

Background: Anti-cancer drug response has been urgently required for individualized therapy. Measurements with wet experiments are costly and time-consuming. Artificial intelligence-based models are currently available for predicting drug response but still have challenges in prediction accuracy. Objective: Construct a model to predict drug response values for unknown cell lines and analyze drug potential association properties in sparse data. Methods: Propose a neural matrix factorization (NeuMF) framework to help predict the unknown responses of cell lines to drugs. The model uses a deep neural network to figure out latent variables for both drugs and cell lines. In NeuMF, the inputs and the parameters of the multi-layer neural network are simultaneously optimized by gradient descent to minimize the reconstruction errors between the predicted and natural values of the observed entries. Then the unknown entries can be readily recovered by propagating the latent variables to the output layer. Results: Experiments on the Cancer Cell Line Encyclopedia (CCLE) dataset and Genomics of Drug Sensitivity in Cancer (GDSC) dataset compare NeuMF with the other three state-of-the-art methods. NeuMF reduces constructing drug or cell line similarity and mines the response matrix itself for correlations in the network, avoiding the inclusion of redundant noise. NeuMF obtained drug averaged PCC_sr of 0.83 and 0.84 on both datasets. It demonstrates that NeuMF substantially improves the prediction. Some essential parameters in NeuMF, such as the strategy of global effect removal and the scales of the input layer, are also discussed. Finally, case studies have shown that NeuMF can better learn the latent characteristics of drugs, e.g., Irinotecan and Topotecan are found to act on the same pathway TOP1. The conclusions are in line with some existing biological findings. Conclusion: NeuMF achieves better prediction accuracy than existing models, and its output is biologically interpretable. NeuMF also helps in analyzing the correlations between drugs.

DBDNMF: A Dual Branch Deep Neural Matrix Factorization method for drug response prediction

“…It has been shown that a deep neural network can generate better representations of arbitrary nonlinear relations among input entries and outperforms single-layer structures [34]. Through the deep neural network structure, the nonlinear and linear terms are approximately expressed as ( 9) and (10), respectively. f ðdÞ ¼ s ðLþ1Þ ðW ðLÞ c s ðLÞ ðW ðLÞ c ð� � � s ð1Þ ðW ð1Þ…”

“…By solving (6), the optimized f(�), C and D can be achieved. Thus, based on (10), C is approximated by the product of the linear hidden layer weights. In a broad sense, C also represents the latent factor matrix of cell line learned in the DBDNMF.…”

“…Therefore, the loss function of learning the latent representation of drugs (D) and the PLOS COMPUTATIONAL BIOLOGY network parameters can be recast as follows by replacing ( 9) and (10).…”

“…Based on the idea of subspace clustering, Zhang et al proposed a new self-expression matrix completion model (SEMCM) aimed at improving the prediction performance of drug response prediction [ 9 ]. Liu et al proposed a Neural Matrix Factorization (NeuMF) framework, which used a deep neural network to figure out drug and cell lines’ latent variables to help predict the unknown responses of cell lines to drugs [ 10 ]. These MF-based methods have shown good predictive impact by providing linear reconstructions of the observed data.…”