SMRT: Randomized Data Transformation for Cancer Subtyping and Big Data Analysis

Randomized singular value decomposition for integrative subtype analysis of 'omics data' using non-negative matrix factorization

Integration of multiple 'omics datasets for differentiating cancer subtypes is a powerful technic that leverages the consistent and complementary information across multi-omics data. Matrix factorization is a common technique used in integrative clustering for identifying latent subtype structure across multi-omics data. High dimensionality of the omics data and long computation time have been common challenges of clustering methods. In order to address the challenges, we propose randomized singular value decomposition (RSVD) for integrative clustering using Non-negative Matrix Factorization: intNMF-rsvd. The method utilizes RSVD to reduce the dimensionality by projecting the data into eigen vector space with user specified lower rank. Then, clustering analysis is carried out by estimating common basis matrix across the projected multi-omics datasets. The performance of the proposed method was assessed using the simulated datasets and compared with six state-of-the-art integrative clustering methods using real-life datasets from The Cancer Genome Atlas Study. intNMF-rsvd was found working efficiently and competitively as compared to standard intNMF and other multi-omics clustering methods. Most importantly, intNMF-rsvd can handle large number of features and significantly reduce the computation time. The identified subtypes can be utilized for further clinical association studies to understand the etiology of the disease.

Integrative analysis of m3C associated genes reveals METTL2A as a potential oncogene in breast Cancer

RNA methylation modifications, especially m6A mRNA modification, are known to be extensively involved in tumor development. However, the relationship between N3-methylcytidine (m3C) related genes and tumorigenesis has rarely been studied. In this research, we found that m3C-related genes were expressed at different levels and affected patients’ prognosis across multiple cancer types from The Cancer Genome Atlas and multi-omics levels. Importantly, methyltransferase-like proteins 2A (METTL2A) had a high amplification frequency (~ 7%) in patients with breast invasive carcinoma (BRCA), and its overexpression was an independent predictor of poor overall survival. Enrichment analysis of associated genes revealed that METTL2A may activate DNA synthesis and cell proliferation pathways in BRCA cells. Through drug sensitivity analysis, Trifluridine, PD407824, and Taselisib were shown to be effective drugs for METTL2A-positive BRCA patients. Overall, our research conducts a holistic view of the expression level and prognostic signature of m3C-related genes with multiple malignancies. Importantly, METTL2A has been intensely explored as a potential oncogene in BRCA, to aid the development of potential drug agents for precision therapy in breast cancer patients.

Crosstalk Between Metabolism and Immune Activity Reveals Four Subtypes With Therapeutic Implications in Clear Cell Renal Cell Carcinoma

Clear cell renal cell carcinoma (ccRCC) is characterized by metabolic dysregulation and distinct immunological signatures. The interplay between metabolic and immune processes in the tumor microenvironment (TME) causes the complexity and heterogeneity of immunotherapy responses observed during ccRCC treatment. Herein, we initially identified two distinct metabolic subtypes (C1 and C2 subtypes) and immune subtypes (I1 and I2 subtypes) based on the occurrence of differentially expressed metabolism-related prognostic genes and immune-related components. Notably, we observed that immune regulators with upregulated expression actively participated in multiple metabolic pathways. Therefore, we further delineated four immunometabolism-based ccRCC subtypes (M1, M2, M3, and M4 subtypes) according to the results of the above classification. Generally, we found that high metabolic activity could suppress immune infiltration. Immunometabolism subtype classification was associated with immunotherapy response, with patients possessing the immune-inflamed, metabolic-desert subtype (M3 subtype) that benefits the most from immunotherapy. Moreover, differences in the shifts in the immunometabolism subtype after immunotherapy were observed in the responder and non-responder groups, with patients from the responder group transferring to subtypes with immune-inflamed characteristics and less active metabolic activity (M3 or M4 subtype). Immunometabolism subtypes could also serve as biomarkers for predicting immunotherapy response. To decipher the genomic and epigenomic features of the four subtypes, we analyzed multiomics data, including miRNA expression, DNA methylation status, copy number variations occurrence, and somatic mutation profiles. Patients with the M2 subtype possessed the highest VHL gene mutation rates and were more likely to be sensitive to sunitinib therapy. Moreover, we developed non-invasive radiomic models to reveal the status of immune activity and metabolism. In addition, we constructed a radiomic prognostic score (PRS) for predicting ccRCC survival based on the seven radiomic features. PRS was further demonstrated to be closely linked to immunometabolism subtype classification, immune score, and tumor mutation burden. The prognostic value of the PRS and the association of the PRS with immune activity and metabolism were validated in our cohort. Overall, our study established four immunometabolism subtypes, thereby revealing the crosstalk between immune and metabolic activities and providing new insights into personal therapy selection.

A novel method for single-cell data imputation using subspace regression

Recent advances in biochemistry and single-cell RNA sequencing (scRNA-seq) have allowed us to monitor the biological systems at the single-cell resolution. However, the low capture of mRNA material within individual cells often leads to inaccurate quantification of genetic material. Consequently, a significant amount of expression values are reported as missing, which are often referred to as dropouts. To overcome this challenge, we develop a novel imputation method, named single-cell Imputation via Subspace Regression (scISR), that can reliably recover the dropout values of scRNA-seq data. The scISR method first uses a hypothesis-testing technique to identify zero-valued entries that are most likely affected by dropout events and then estimates the dropout values using a subspace regression model. Our comprehensive evaluation using 25 publicly available scRNA-seq datasets and various simulation scenarios against five state-of-the-art methods demonstrates that scISR is better than other imputation methods in recovering scRNA-seq expression profiles via imputation. scISR consistently improves the quality of cluster analysis regardless of dropout rates, normalization techniques, and quantification schemes. The source code of scISR can be found on GitHub at https://github.com/duct317/scISR.