CEPICS: A Comparison and Evaluation Platform for Integration Methods in Cancer Subtyping

Abstract

Cancer subtypes can improve our understanding of cancer, and suggest more precise treatment for patients. Multi-omics molecular data can characterize cancers at different levels. Up to now, many computational methods that integrate multi-omics data for cancer subtyping have been proposed. However, there are no consistent criteria to evaluate the integration methods due to the lack of gold standards (e.g., the number of subtypes in a specific cancer). Since comprehensive evaluation and comparison between different methods serves as a useful tool or guideline for users to select an optimal method for their own purpose, we develop a scalable platform, CEPICS, for comprehensively evaluating and comparing multi-omics data integration methods in cancer subtyping. Given a user-specified maximum number of subtypes, k-max, CEPICS provides (1) cancer subtyping results using up to five built-in state-of-the-art integration methods under the number of subtypes from two to k-max, (2) a report including the evaluation of each user-selected method and comparisons across them using clustering performance metrics and clinical survival analysis, and (3) an overall analysis of subtyping results by different methods representing a robust cancer subtype prediction for samples. Furthermore, users can upload subtyping results of their own methods to compare with the built-in methods. CEPICS is implemented as an R package and is freely available at .

Evaluation and comparison of multi-omics data integration methods for cancer subtyping

“…We also did some improvements to these packages, including the parallelization of LRAcluster, running iClusterBayes in parallel without operating system limitation, and removing PFA's limitation on the number of data types it could handle. Details of these technical improvements can be found in our recent paper [6] (https://github.com/GaoLabXDU/CEPICS). Since the LRAcluster, PFA, and MultiNMF package only output an integrated sample-feature matrix rather than clustering (subtyping) results, we used the K-means clustering with 300 iterations on the sample-feature matrices to obtain stable subtyping results.…”

“…The first problem is about how to compare the performance among these methods and the second problem is related to the selection of available data types to integrate in order to achieve the best possible results. The first problem is due to the lack of gold standards and consistent performic criteria [6], and the fact that different datasets and evaluation metrics were used when different methods were proposed. To understand and demonstrate the crucial need for addressing the second problem of data type selection, we surveyed 58 integration methods for cancer subtyping proposed from 2009 to 2019, and the result is summarized in Fig 1 where gene expression is treated as the same as mRNA expression and miRNA expression is placed into the group of epigenome based on observations from [7].…”

SSIG: Single‐Sample Information Gain Model for Integrating Multi‐Omics Data to Identify Cancer Subtypes

“…Considering the lack of a gold standard in the identification of cancer subtypes at present [ 14 ] , we download two different types of cancer data, BRCA and KIRC, from the TCGA database as two types of real labels for the samples to test the performance of different methods for the identification of cancer subtypes.…”

“…SNF, ANF, SKF and SRF are four middle integration methods to obtain the sample similarity network based on sample similarity fusion [ 13 ] , and then conduct clustering analysis according to the obtained similarity network, so the user needs to specify the number of clusters [ 14 ] . In this paper, we set the number of clusters to 2, 3, 4, 5 and analyzed them respectively.…”

Multi-omics peripheral and core regions of cancer

“…Based on multi-omics data, Ding et al found (1) somatic driver mutations, germline pathogenic variants, and their interactions in tumours; (2) the tumour genome and epigenome’s influence on the transcriptome and proteome; and (3) the relationship between the tumour and the microenvironment 45 . Using multi-omics technologies, Bhattacharya et al performed transcriptome-wide association studies 46 and Duan et al 47 analysed cancer subtypes. Shi et al 48 have developed a novel algorithm, Iterative Clique Enumeration (ICE), for identifying relatively independent maximal cliques as co-expression modules and a module-based approach to the analysis of gene expression data.…”