A new advanced in silico drug discovery method for novel coronavirus (SARS-CoV-2) with tensor decomposition-based unsupervised feature extraction

Abstract

Background: COVID-19 is a critical pandemic that has affected human communities worldwide, and there is an urgent need to develop effective drugs. Although there are a large number of candidate drug compounds that may be useful for treating COVID-19, the evaluation of these drugs is time-consuming and costly. Thus, screening to identify potentially effective drugs prior to experimental validation is necessary. Method: In this study, we applied the recently proposed method tensor decomposition (TD)-based unsupervised feature extraction (FE) to gene expression profiles of multiple lung cancer cell lines infected with severe acute respiratory syndrome coronavirus 2. We identified drug candidate compounds that significantly altered the expression of the 163 genes selected by TD-based unsupervised FE. Results: Numerous drugs were successfully screened, including many known antiviral drug compounds such as C646, chelerythrine chloride, canertinib, BX-795, sorafenib, sorafenib, QL-X-138, radicicol, A-443654, CGP-60474, alvocidib, mitoxantrone, QL-XII-47, geldanamycin, fluticasone, atorvastatin, quercetin, motexafin gadolinium, trovafloxacin, doxycycline, meloxicam, gentamicin, and dibromochloromethane. The screen also identified ivermectin, which was first identified as an anti-parasite drug and recently the drug was included in clinical trials for SARS-CoV-2. Conclusions: The drugs screened using our strategy may be effective candidates for treating patients with COVID-19.

Mathematical formulation and application of kernel tensor decomposition based unsupervised feature extraction

“…The application of TD based unsupervised FE to problems in genomic science yielded satisfactory results even when conventional feature selection methods based upon statistical tests failed (Taguchi, 2020; Taguchi and Turki, 2020; Ng and Taguchi, 2020). Nevertheless, TD based unsupervised FE involve certain limitations.…”

“…In particular, the TD based unsupervised FE could predict many promising drugs including ivermectin, the clinical trials using which have been recently initiated. Herein, we briefly summarize the process implemented in the previous work (Taguchi and Turki, 2020) to enable a comparative analysis of the results of the KTD based unsupervised FE with the previous results. The gene expression profiles was formatted as a tensor, x ijkm ∈ ℝ N ×5×2×3 , which indicated whether the gene expression of the i th gene of the j th cell line infected ( k = 1) or not infected ( k = 2, control) considering three biological replicates.…”

“…In particular, we compared the selected 163 genes with gold standard human proteins that are known to interact with the SARS-CoV-2 proteins during infection (Gordon et al, 2020), following the procedure described in the previous study (Taguchi and Turki, 2020) (Table 3). Although the 163 genes selected by the KTD based unsupervised FE are highly coincident with the gold standard SARS-CoV-2 interacting human proteins, the results in Table 3 are slightly less coincident with the results for the gold standard SARS-CoV-2 interacting human proteins compared to the results presented in Table S33 (Taguchi and Turki, 2020) corresponding to the application of the TD based unsupervised FE (Taguchi and Turki, 2020). Notably, only one protein group (SARS-CoV2 N) among the 27 protein groups in Table 3 exhibits a larger odds ratio in Table 3 (corresponding to the KTD based unsupervised FE) than that in Table S33 (corresponding to the TD based unsupervised FE (Taguchi and Turki, 2020)).…”

Projection in genomic analysis: A theoretical basis to rationalize tensor decomposition and principal component analysis as feature selection tools

“…To explain why PCA- and TD-based unsupervised FE work rather well, we consider two recent works 9,10 , in which the superiority of PCA- and/or TD-based unsupervised FE over conventional statistical methods was shown; in these studies, conventional statistical test-based methods failed to select a reasonable number of genes whereas TD-based unsupervised FE successfully selected a biologically reasonable restricted number of genes.…”

“…In this paper, we try to understand why PCA-based unsupervised FE and TD-based unsupervised FE 3 are effective in feature selection based on projection strategy. We consider the cases SARS-CoV-2 infection problem 9 as well as biomarker identification of kidney cancer 10 ; in these studies, despite unsuccessful results obtained by conventional feature selection based on statistical tests, TD-based unsupervised FE identified biologically reasonable genes (for more details about how PCA- and TD-based unsupervised FE are superior to statistical test-based feature selection tools in these specific examples, see these previous studies 9,10) .…”