PATHOME-Drug: a subpathway-based polypharmacology drug-repositioning method.

MOTIVATION: Drug repositioning reveals novel indications for existing drugs and in particular, diseases with no available drugs. Diverse computational drug repositioning methods have been proposed by measuring either drug-treated gene expression signatures or the proximity of drug targets and disease proteins found in prior networks. However, these methods do not explain which signaling subparts allow potential drugs to be selected, and do not consider polypharmacology, i.e. multiple targets of a known drug, in specific subparts. RESULTS: Here, to address the limitations, we developed a subpathway-based polypharmacology drug repositioning method, PATHOME-Drug, based on drug-associated transcriptomes. Specifically, this tool locates subparts of signaling cascading related to phenotype changes (e.g. disease status changes), and identifies existing approved drugs such that their multiple targets are enriched in the subparts. We show that our method demonstrated better performance for detecting signaling context and specific drugs/compounds, compared to WebGestalt and clusterProfiler, for both real biological and simulated datasets. We believe that our tool can successfully address the current shortage of targeted therapy agents. AVAILABILITY AND IMPLEMENTATION: The web-service is available at http://statgen.snu.ac.kr/software/pathome. The source codes and data are available at https://github.com/labnams/pathome-drug. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

A comprehensive evaluation of regression-based drug responsiveness prediction models, using cell viability inhibitory concentrations (IC50 values).

MOTIVATION: Predicting drug response is critical for precision medicine. Diverse methods have predicted drug responsiveness, as measured by the half-maximal drug inhibitory concentration (IC50), in cultured cells. Although IC50s are continuous, traditional prediction models have dealt mainly with binary classification of responsiveness. However, since there are few regression-based IC50 predictions, comprehensive evaluations of regression-based IC50 prediction models, including machine learning (ML) and deep learning (DL), for diverse data types and dataset sizes, have not been addressed. RESULTS: Here, we constructed 11 input data settings, including multi-omics settings, with varying dataset sizes, then evaluated the performance of regression-based ML and DL models to predict IC50s. DL models considered two convolutional neural network architectures: CDRScan and residual neural network (ResNet). ResNet was introduced in regression-based DL models for predicting drug response for the first time. As a result, DL models performed better than ML models in all the settings. Also, ResNet performed better than or comparable to CDRScan and ML models in all settings. AVAILABILITY AND IMPLEMENTATION: The data underlying this article are available in GitHub at https://github.com/labnams/IC50evaluation. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Plasma Aldo-Keto Reductase Family 1 Member B10 as a Biomarker Performs Well in the Diagnosis of Nonalcoholic Steatohepatitis and Fibrosis.

We found several blood biomarkers through computational secretome analyses, including aldo-keto reductase family 1 member B10 (AKR1B10), which reflected the progression of nonalcoholic fatty liver disease (NAFLD). After confirming that hepatic AKR1B10 reflected the progression of NAFLD in a subgroup with NAFLD, we evaluated the diagnostic accuracy of plasma AKR1B10 and other biomarkers for the diagnosis of nonalcoholic steatohepatitis (NASH) and fibrosis in replication cohort. We enrolled healthy control subjects and patients with biopsy-proven NAFLD (n = 102) and evaluated the performance of various diagnostic markers. Plasma AKR1B10 performed well in the diagnosis of NASH with an area under the receiver operating characteristic (AUROC) curve of 0.834 and a cutoff value of 1078.2 pg/mL, as well as advanced fibrosis (AUROC curve value of 0.914 and cutoff level 1078.2 pg/mL), with further improvement in combination with C3. When we monitored a subgroup of obese patients who underwent bariatric surgery (n = 35), plasma AKR1B10 decreased dramatically, and 40.0% of patients with NASH at baseline showed a decrease in plasma AKR1B10 levels to below the cutoff level after the surgery. In an independent validation study, we proved that plasma AKR1B10 was a specific biomarker of NAFLD progression across varying degrees of renal dysfunction. Despite perfect correlation between plasma and serum levels of AKR1B10 in paired sample analysis, its serum level was 1.4-fold higher than that in plasma. Plasma AKR1B10 alone and in combination with C3 could be a useful noninvasive biomarker for the diagnosis of NASH and hepatic fibrosis.

A Deep Learning Model for Cell Growth Inhibition IC50 Prediction and Its Application for Gastric Cancer Patients.

Heterogeneity in intratumoral cancers leads to discrepancies in drug responsiveness, due to diverse genomics profiles. Thus, prediction of drug responsiveness is critical in precision medicine. So far, in drug responsiveness prediction, drugs' molecular "fingerprints", along with mutation statuses, have not been considered. Here, we constructed a 1-dimensional convolution neural network model, DeepIC50, to predict three drug responsiveness classes, based on 27,756 features including mutation statuses and various drug molecular fingerprints. As a result, DeepIC50 showed better cell viability IC50 prediction accuracy in pan-cancer cell lines over two independent cancer cell line datasets. Gastric cancer (GC) is not only one of the lethal cancer types in East Asia, but also a heterogeneous cancer type. Currently approved targeted therapies in GC are only trastuzumab and ramucirumab. Responsive GC patients for the drugs are limited, and more drugs should be developed in GC. Due to the importance of GC, we applied DeepIC50 to a real GC patient dataset. Drug responsiveness prediction in the patient dataset by DeepIC50, when compared to the other models, were comparable to responsiveness observed in GC cell lines. DeepIC50 could possibly accurately predict drug responsiveness, to new compounds, in diverse cancer cell lines, in the drug discovery process.

Zingerone Suppresses Tumor Development through Decreasing Cyclin D1 Expression and Inducing Mitotic Arrest.

Cancer cells undergo uncontrolled proliferation resulting from aberrant activity of various cell-cycle proteins. Therefore, despite recent advances in intensive chemotherapy, it is difficult to cure cancer completely. Recently, cell-cycle regulators became attractive targets in cancer therapy. Zingerone, a phenolic compound isolated from ginger, is a nontoxic and inexpensive compound with varied pharmacological activities. In this study, the therapeutic effect of zingerone as an anti-mitotic agent in human neuroblastoma cells was investigated. Following treatment of BE(2)-M17 cells with zingerone, we performed a 3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide (MTT) assay and colony-formation assay to evaluate cellular proliferation, in addition to immunofluorescence cytochemistry and flow cytometry to examine the mitotic cells. The association of gene expression with tumor stage and survival was analyzed. Furthermore, to examine the anti-cancer effect of zingerone, we applied a BALB/c mouse-tumor model using a BALB/c-derived adenocarcinoma cell line. In human neuroblastoma cells, zingerone inhibited cellular viability and survival. Moreover, the number of mitotic cells, particularly those in prometaphase, increased in zingerone-treated neuroblastoma cells. Regarding specific molecular mechanisms, zingerone decreased cyclin D1 expression and induced the cleavage of caspase-3 and poly (ADP-ribose) polymerase 1 (PARP-1). The decrease in cyclin D1 and increase in histone H3 phosphorylated (p)-Ser10 were confirmed by immunohistochemistry in tumor tissues administered with zingerone. These results suggest that zingerone induces mitotic arrest followed by inhibition of growth of neuroblastoma cells. Collectively, zingerone may be a potential therapeutic drug for human cancers, including neuroblastoma.

miRDM-rfGA: Genetic algorithm-based identification of a miRNA set for detecting type 2 diabetes.

BACKGROUND: Type 2 diabetes mellitus (T2DM) affects approximately 451 million adults globally. In this study, we identified the optimal combination of marker candidates for detecting T2DM using miRNA-Seq data from 95 samples including T2DM and healthy individuals. METHODS: We utilized the genetic algorithm (GA) in the discovery of an optimal miRNA biomarker set. We discovered miRNA subsets consisting of three miRNAs for detecting T2DM by random forest-based GA (miRDM-rfGA) as a feature selection algorithm and created six GA parameter settings and three settings using traditional feature selection methods (F-test and Lasso). We then evaluated the prediction performance to detect T2DM in the miRNA subsets derived from each setting. RESULTS: The miRNA subset in setting 5 using miRDM-rfGA performed the best in detecting T2DM (mean AUROC = 0.92). Target mRNA identification and functional enrichment analysis of the best miRNA subset (hsa-miR-125b-5p, hsa-miR-7-5p, and hsa-let-7b-5p) validated that this combination was involved in T2DM. We also confirmed that the targeted genes were negatively correlated with the clinical variables related to T2DM in the BxD mouse genetic reference population database. CONCLUSIONS: Using GA in miRNA-Seq data, we identified the optimal miRNA biomarker set for T2DM detection. GA can be a useful tool for biomarker discovery and drug-target identification.

A performance evaluation of drug response prediction models for individual drugs.

Drug response prediction is important to establish personalized medicine for cancer therapy. Model construction for predicting drug response (i.e., cell viability half-maximal inhibitory concentration [IC50]) of an individual drug by inputting pharmacogenomics in disease models remains critical. Machine learning (ML) has been predominantly applied for prediction, despite the advent of deep learning (DL). Moreover, whether DL or traditional ML models are superior for predicting cell viability IC50s has to be established. Herein, we constructed ML and DL drug response prediction models for 24 individual drugs and compared the performance of the models by employing gene expression and mutation profiles of cancer cell lines as input. We observed no significant difference in drug response prediction performance between DL and ML models for 24 drugs [root mean squared error (RMSE) ranging from 0.284 to 3.563 for DL and from 0.274 to 2.697 for ML; R2 ranging from -7.405 to 0.331 for DL and from -8.113 to 0.470 for ML]. Among the 24 individual drugs, the ridge model of panobinostat exhibited the best performance (R2 0.470 and RMSE 0.623). Thus, we selected the ridge model of panobinostat for further application of explainable artificial intelligence (XAI). Using XAI, we further identified important genomic features for panobinostat response prediction in the ridge model, suggesting the genomic features of 22 genes. Based on our findings, results for an individual drug employing both DL and ML models were comparable. Our study confirms the applicability of drug response prediction models for individual drugs.

Machine learning-derived gut microbiome signature predicts fatty liver disease in the presence of insulin resistance.

A simple predictive biomarker for fatty liver disease is required for individuals with insulin resistance. Here, we developed a supervised machine learning-based classifier for fatty liver disease using fecal 16S rDNA sequencing data. Based on the Kangbuk Samsung Hospital cohort (n = 777), we generated a random forest classifier to predict fatty liver diseases in individuals with or without insulin resistance (n = 166 and n = 611, respectively). The model performance was evaluated based on metrics, including accuracy, area under receiver operating curve (AUROC), kappa, and F1-score. The developed classifier for fatty liver diseases performed better in individuals with insulin resistance (AUROC = 0.77). We further optimized the classifiers using genetic algorithm. The improved classifier for insulin resistance, consisting of ten microbial genera, presented an advanced classification (AUROC = 0.93), whereas the improved classifier for insulin-sensitive individuals failed to distinguish participants with fatty liver diseases from the healthy. The classifier for individuals with insulin resistance was comparable or superior to previous methods predicting fatty liver diseases (accuracy = 0.83, kappa = 0.50, F1-score = 0.89), such as the fatty liver index. We identified the ten genera as a core set from the human gut microbiome, which could be a diagnostic biomarker of fatty liver diseases for insulin resistant individuals. Collectively, these findings indicate that the machine learning classifier for fatty liver diseases in the presence of insulin resistance is comparable or superior to commonly used methods.

A Comparison of the Transcriptomes of Cowpeas in Response to Two Different Ionizing Radiations.

In this study, gene expression changes in cowpea plants irradiated by two different types of radiation: proton-beams and gamma-rays were investigated. Seeds of the Okdang cultivar were exposed to 100, 200, and 300 Gy of gamma-rays and proton-beams. In transcriptome analysis, the 32, 75, and 69 differentially expressed genes (DEGs) at each dose of gamma-ray irradiation compared with that of the control were identified. A total of eight genes were commonly up-regulated for all gamma-ray doses. However, there were no down-regulated genes. In contrast, 168, 434, and 387 DEGs were identified for each dose of proton-beam irradiation compared with that of the control. A total of 61 DEGs were commonly up-regulated for all proton-beam doses. As a result of GO and KEGG analysis, the ranks of functional categories according to the number of DEGs were not the same in both treatments and were more diverse in terms of pathways in the proton-beam treatments than gamma-ray treatments. The number of genes related to defense, photosynthesis, reactive oxygen species (ROS), plant hormones, and transcription factors (TF) that were up-/down-regulated was higher in the proton beam treatment than that in gamma ray treatment. Proton-beam treatment had a distinct mutation spectrum and gene expression pattern compared to that of gamma-ray treatment. These results provide important information on the mechanism for gene regulation in response to two ionizing radiations in cowpeas.

Deep learning for stage prediction in neuroblastoma using gene expression data.

Neuroblastoma is a major cause of cancer death in early childhood, and its timely and correct diagnosis is critical. Gene expression datasets have recently been considered as a powerful tool for cancer diagnosis and subtype classification. However, no attempts have yet been made to apply deep learning using gene expression to neuroblastoma classification, although deep learning has been applied to cancer diagnosis using image data. Taking the International Neuroblastoma Staging System stages as multiple classes, we designed a deep neural network using the gene expression patterns and stages of neuroblastoma patients. Despite a small patient population (n = 280), stage 1 and 4 patients were well distinguished. If it is possible to replicate this approach in a larger population, deep learning could play an important role in neuroblastoma staging.