Dual Representation Learning for Predicting Drug-side Effect Frequency using Protein Target Information.

Knowledge of unintended effects of drugs is critical in assessing the risk of treatment and in drug repurposing. Although numerous existing studies predict drug-side effect presence, only four of them predict the frequency of the side effects. Unfortunately, current prediction methods (1) do not utilize drug targets, (2) do not predict well for unseen drugs, and (3) do not use multiple heterogeneous drug features. We propose a novel deep learning-based drug-side effect frequency prediction model. Our model utilized heterogeneous features such as target protein information as well as molecular graph, fingerprints, and chemical similarity to create drug embeddings simultaneously. Furthermore, the model represents drugs and side effects into a common vector space, learning the dual representation vectors of drugs and side effects, respectively. We also extended the predictive power of our model to compensate for the drugs without clear target proteins using the Adaboost method. We achieved state-of-the-art performance over the existing methods in predicting side effect frequencies, especially for unseen drugs. Ablation studies show that our model effectively combines and utilizes heterogeneous features of drugs. Moreover, we observed that, when the target information given, drugs with explicit targets resulted in better prediction than the drugs without explicit targets. The implementation is available at https://github.com/eskendrian/sider.

Improved drug response prediction by drug target data integration via network-based profiling.

Drug response prediction (DRP) is important for precision medicine to predict how a patient would react to a drug before administration. Existing studies take the cell line transcriptome data, and the chemical structure of drugs as input and predict drug response as IC50 or AUC values. Intuitively, use of drug target interaction (DTI) information can be useful for DRP. However, use of DTI is difficult because existing drug response database such as CCLE and GDSC do not have information about transcriptome after drug treatment. Although transcriptome after drug treatment is not available, if we can compute the perturbation effects by the pharmacologic modulation of target gene, we can utilize the DTI information in CCLE and GDSC. In this study, we proposed a framework that can improve existing deep learning-based DRP models by effectively utilizing drug target information. Our framework includes NetGP, a module to compute gene perturbation scores by the network propagation technique on a network. NetGP produces genes in a ranked list in terms of gene perturbation scores and the ranked genes are input to a multi-layer perceptron to generate a fixed dimension vector for the integration with existing DRP models. This integration is done in a model-agnostic way so that any existing DRP tool can be incorporated. As a result, our framework boosts the performance of existing DRP models, in 64 of 72 comparisons. The performance gains are larger especially for test scenarios with samples with unseen drugs by large margins up to 34% in Pearson's correlation coefficient.

Hepatic GSK3ß-Dependent CRY1 Degradation Contributes to Diabetic Hyperglycemia.

Excessive hepatic glucose production (HGP) is a key factor promoting hyperglycemia in diabetes. Hepatic cryptochrome 1 (CRY1) plays an important role in maintaining glucose homeostasis by suppressing forkhead box O1 (FOXO1)-mediated HGP. Although downregulation of hepatic CRY1 appears to be associated with increased HGP, the mechanism(s) by which hepatic CRY1 dysregulation confers hyperglycemia in subjects with diabetes is largely unknown. In this study, we demonstrate that a reduction in hepatic CRY1 protein is stimulated by elevated E3 ligase F-box and leucine-rich repeat protein 3 (FBXL3)-dependent proteasomal degradation in diabetic mice. In addition, we found that GSK3ß-induced CRY1 phosphorylation potentiates FBXL3-dependent CRY1 degradation in the liver. Accordingly, in diabetic mice, GSK3ß inhibitors effectively decreased HGP by facilitating the effect of CRY1-mediated FOXO1 degradation on glucose metabolism. Collectively, these data suggest that tight regulation of hepatic CRY1 protein stability is crucial for maintaining systemic glucose homeostasis.

AutoCoV: tracking the early spread of COVID-19 in terms of the spatial and temporal patterns from embedding space by K-mer based deep learning.

BACKGROUND: The widely spreading coronavirus disease (COVID-19) has three major spreading properties: pathogenic mutations, spatial, and temporal propagation patterns. We know the spread of the virus geographically and temporally in terms of statistics, i.e., the number of patients. However, we are yet to understand the spread at the level of individual patients. As of March 2021, COVID-19 is wide-spread all over the world with new genetic variants. One important question is to track the early spreading patterns of COVID-19 until the virus has got spread all over the world. RESULTS: In this work, we proposed AutoCoV, a deep learning method with multiple loss object, that can track the early spread of COVID-19 in terms of spatial and temporal patterns until the disease is fully spread over the world in July 2020. Performances in learning spatial or temporal patterns were measured with two clustering measures and one classification measure. For annotated SARS-CoV-2 sequences from the National Center for Biotechnology Information (NCBI), AutoCoV outperformed seven baseline methods in our experiments for learning either spatial or temporal patterns. For spatial patterns, AutoCoV had at least 1.7-fold higher clustering performances and an F1 score of 88.1%. For temporal patterns, AutoCoV had at least 1.6-fold higher clustering performances and an F1 score of 76.1%. Furthermore, AutoCoV demonstrated the robustness of the embedding space with an independent dataset, Global Initiative for Sharing All Influenza Data (GISAID). CONCLUSIONS: In summary, AutoCoV learns geographic and temporal spreading patterns successfully in experiments on NCBI and GISAID datasets and is the first of its kind that learns virus spreading patterns from the genome sequences, to the best of our knowledge. We expect that this type of embedding method will be helpful in characterizing fast-evolving pandemics.

MLDEG: A Machine Learning Approach to Identify Differentially Expressed Genes Using Network Property and Network Propagation.

MOTIVATION: Identifying differentially expressed genes (DEGs) in transcriptome data is a very important task. However, performances of existing DEG methods vary significantly for data sets measured in different conditions and no single statistical or machine learning model for DEG detection perform consistently well for data sets of different traits. In addition, setting a cutoff value for the significance of differential expressions is one of confounding factors to determine DEGs. RESULTS: We address these problems by developing an ensemble model that refines the heterogeneous and inconsistent results of the existing methods by taking accounts into network information such as network propagation and network property. DEG candidates that are predicted with weak evidence by the existing tools are re-classified by our proposed ensemble model for the transcriptome data. Tested on 10 RNA-seq datasets downloaded from gene expression omnibus (GEO), our method showed excellent performance of winning the first place in detecting ground truth (GT) genes in eight datasets and find almost all GT genes in six datasets. On the other hand, performances of all existing methods varied significantly for the 10 data sets. Because of the design principle, our method can accommodate any new DEG methods naturally. AVAILABILITY: The source code of our method is available at https://github.com/jihmoon/MLDEG.

Impact of mutations in DNA methylation modification genes on genome-wide methylation landscapes and downstream gene activations in pan-cancer.

BACKGROUND: In cancer, mutations of DNA methylation modification genes have crucial roles for epigenetic modifications genome-wide, which lead to the activation or suppression of important genes including tumor suppressor genes. Mutations on the epigenetic modifiers could affect the enzyme activity, which would result in the difference in genome-wide methylation profiles and, activation of downstream genes. Therefore, we investigated the effect of mutations on DNA methylation modification genes such as DNMT1, DNMT3A, MBD1, MBD4, TET1, TET2 and TET3 through a pan-cancer analysis. METHODS: First, we investigated the effect of mutations in DNA methylation modification genes on genome-wide methylation profiles. We collected 3,644 samples that have both of mRNA and methylation data from 12 major cancer types in The Cancer Genome Atlas (TCGA). The samples were divided into two groups according to the mutational signature. Differentially methylated regions (DMR) that overlapped with the promoter region were selected using minfi and differentially expressed genes (DEG) were identified using EBSeq. By integrating the DMR and DEG results, we constructed a comprehensive DNA methylome profiles on a pan-cancer scale. Second, we investigated the effect of DNA methylations in the promoter regions on downstream genes by comparing the two groups of samples in 11 cancer types. To investigate the effects of promoter methylation on downstream gene activations, we performed clustering analysis of DEGs. Among the DEGs, we selected highly correlated gene set that had differentially methylated promoter regions using graph based sub-network clustering methods. RESULTS: We chose an up-regulated DEGs cluster where had hypomethylated promoter in acute myeloid leukemia (LAML) and another down-regulated DEGs cluster where had hypermethylated promoter in colon adenocarcinoma (COAD). To rule out effects of gene regulation by transcription factor (TF), if differentially expressed TFs bound to the promoter of DEGs, that DEGs did not included to the gene set that effected by DNA methylation modifiers. Consequently, we identified 54 hypomethylated promoter DMR up-regulated DEGs in LAML and 45 hypermethylated promoter DMR down-regulated DEGs in COAD. CONCLUSIONS: Our study on DNA methylation modification genes in mutated vs. non-mutated groups could provide useful insight into the epigenetic regulation of DEGs in cancer.

PropaNet: Time-Varying Condition-Specific Transcriptional Network Construction by Network Propagation.

Transcription factor (TF) has a significant influence on the state of a cell by regulating multiple down-stream genes. Thus, experimental and computational biologists have made great efforts to construct TF gene networks for regulatory interactions between TFs and their target genes. Now, an important research question is how to utilize TF networks to investigate the response of a plant to stress at the transcription control level using time-series transcriptome data. In this article, we present a new computational network, PropaNet, to investigate dynamics of TF networks from time-series transcriptome data using two state-of-the-art network analysis techniques, influence maximization and network propagation. PropaNet uses the influence maximization technique to produce a ranked list of TFs, in the order of TF that explains differentially expressed genes (DEGs) better at each time point. Then, a network propagation technique is used to select a group of TFs that explains DEGs best as a whole. For the analysis of Arabidopsis time series datasets from AtGenExpress, we used PlantRegMap as a template TF network and performed PropaNet analysis to investigate transcriptional dynamics of Arabidopsis under cold and heat stress. The time varying TF networks showed that Arabidopsis responded to cold and heat stress quite differently. For cold stress, bHLH and bZIP type TFs were the first responding TFs and the cold signal influenced histone variants, various genes involved in cell architecture, osmosis and restructuring of cells. However, the consequences of plants under heat stress were up-regulation of genes related to accelerating differentiation and starting re-differentiation. In terms of energy metabolism, plants under heat stress show elevated metabolic process and resulting in an exhausted status. We believe that PropaNet will be useful for the construction of condition-specific time-varying TF network for time-series data analysis in response to stress. PropaNet is available at http://biohealth.snu.ac.kr/software/PropaNet.