Skip to main content
NCBI home page
Bioinformatics logoLink to Bioinformatics
. 2023 Nov 23;39(12):btad708. doi: 10.1093/bioinformatics/btad708

COMMO: a web server for the identification and analysis of consensus gene modules across multiple methods

Xiaojing Wu 1,2,2, Mingfei Han 3,2, Xinyu Song 4,2, Song He 5, Xiaochen Bo 6,✉, Yunping Zhu 7,8,✉
Editor: Pier Luigi Martelli
PMCID: PMC10713113  PMID: 37995293

Abstract

Summary

A variety of computational methods have been developed to identify functionally related gene modules from genome-wide gene expression profiles. Integrating the results of these methods to identify consensus modules is a promising approach to produce more accurate and robust results. In this application note, we introduce COMMO, the first web server to identify and analyze consensus gene functionally related gene modules from different module detection methods. First, COMMO implements eight state-of-the-art module detection methods and two consensus clustering algorithms. Second, COMMO provides users with mRNA and protein expression data for 33 cancer types from three public databases. Users can also upload their own data for module detection. Third, users can perform functional enrichment and two types of survival analyses on the observed gene modules. Finally, COMMO provides interactive, customizable visualizations and exportable results. With its extensive analysis and interactive capabilities, COMMO offers a user-friendly solution for conducting module-based precision medicine research.

Availability and implementation

COMMO web is available at https://commo.ncpsb.org.cn/, with the source code available on GitHub: https://github.com/Song-xinyu/COMMO/tree/master.

1 Introduction

Functionally related gene modules detection is a critical step in the analysis of genome-wide gene expression profiles generated by high-throughput technologies. These modules can reveal functional relationships between genes and are useful for understanding the complexity of biological systems (Jojic et al. 2013). To date, numerous module detection methods have been developed, including clustering (Kluger et al. 2003), biclustering (Bergmann et al. 2003), matrix decomposition (Wang et al. 2021), and network inference methods (Huynh-Thu et al. 2010). However, many of these existing methods suffer from high parameter dependence and poor robustness (Saelens et al. 2018).

To address this issue, several studies (Brière et al. 2021, Qin et al. 2023) strongly advocated for the integration of modules from various methods to create more accurate and robust results. However, there is a lack of consensus clustering tools or websites that can integrate gene modules identified by different methods. First, existing gene module detection websites, such as ExpressVis (Liu et al. 2022a) and Hiplot (Li et al. 2022), only focus on a single clustering method. Second, existing consensus clustering websites, such as ICM (He et al. 2016) and COMSUC (He et al. 2021), are designed for disease subtyping rather than gene module detection. In summary, there is an urgent need for new tools that use consensus clustering to improve the detection of gene functionally related gene modules.

Here, we present COMMO (COnsensus Molecular MOdules), an interactive web server for biologists to identify and analyze consensus gene functionally related gene modules from different module detection methods. COMMO integrates eight representative module detection methods and two consensus clustering algorithms. Additionally, COMMO provides module-based functional enrichment and survival analysis, which gives researchers practical ways to assess module-disease associations and identify potential biomarkers. COMMO also provides a user-friendly interface and comprehensive result visualization, making it easy for researchers to explore and download results. To evaluate COMMO on clinical cohorts, we applied it to the TCGA breast cancer dataset and identified many well-known pathways and genes associated with breast cancer, as detailed in the Supplementary Materials. In conclusion, we believe that COMMO will be a valuable and convenient platform for combining independently generated clusters into more robust and reliable clusters.

2 Implementation

As shown in Fig. 1A, there are three steps to run COMMO in the web interface: (i) upload expression data to be analyzed, (ii) select module detection and consensus clustering methods, and (iii) explore interactive and exportable results. In the first step, users can upload their own private data or use mRNA and protein expression data across 33 cancer types from three public cancer data sources (TCGA, ICGC, and TARGET).

Figure 1.

Figure 1.

Open in a new tab

(A) The workflow of COMMO web server. (B) The process of COMMO identifying consensus molecular modules. (C) The network architecture of COMMO.

In the second step, COMMO identifies consensus molecular module (CMMs) in three steps (Fig. 1B). First, COMMO allows users to customize the number of clusters. If no cluster number is specified, COMMO will automatically determine the number of clusters using the k-nearest neighbor algorithm (see Supplementary Section S1.1). Second, eight clustering methods are used to identify gene modules, including FLAME (Fuzzy clustering by Local Approximation of Memberships) (Fu and Medico 2007), K-means (Timmerman et al. 2013), SOM (self-organizing mapping) (Wang et al. 2002), spectral clustering (Zhang et al. 2021), Agglomerative (Liu et al. 2022b), Hclust (Bu et al. 2022), NMF (non-negative matrix factorization) (Liefeld et al. 2023), and ICA (independent component analysis) (Hyvärinen 2013). Detailed descriptions of these eight methods can be found in the Supplementary Materials. We selected these methods because they demonstrated superior performance in a previous evaluation study (Saelens et al. 2018) and have reasonable runtimes. Finally, two consensus clustering methods, COCA (Cluster-Of-Cluster-Assignments) (The Cancer Genome Atlas Network 2012, Hoadley et al. 2014, Cabassi and Kirk 2020) and SuperCluster (Kandoth et al. 2013) are used to integrate different clustering results.

In the third step, COMMO provides five types of results, including CMM network maps, CMM heatmaps, module-based functional enrichment analysis, and two types of survival analysis (expression level-based survival analysis and module centrality-based survival analysis) to identify module genes that are associated with cancer patient survival (see Supplementary Methods for details). All graphics generated by the platform can be downloaded for demonstration or publishing, and all data in the table can be downloaded for later use. Notably, to save users time and ensure timely notification even in the event of network interruption, COMMO automatically emails users after a project is completed. Users can then download the result file within 30 days of receiving the email.

The architecture of COMMO is based on the client-server website design (Fig. 1C). The backend server focuses on computation, while the front-end client focuses on interactive visualization. COMMO uses a MongoDB database to store data, and the analysis application is written in R. The web presentation application is implemented in Vue.js. The backend server is deployed in Docker. Network maps are drawn with Cytoscape.js, heatmaps with Highcharts, GO and KEGG enrichment analyses with the ‘enrichGO’ and ‘enrichKEGG’ functions from the ‘clusterProfiler’ R package (Yu et al. 2012), and survival analysis with the ‘survfit’ function from the ‘survival’ R package. COMMO can be deployed locally or remotely on different platforms (including Windows and Linux). The source code is available on GitHub, along with deployment instructions.

3 Conclusion

In conclusion, COMMO is the first web server that can integrate multiple module detection methods to identify CMMs. It also offers extensive functional analysis and visualization capabilities for these modules and enables biologists to identify reliable consensus molecular modules and perform module-based precision medicine research. COMMO is available at https://commo.ncpsb.org.cn/.

Supplementary Material

btad708_Supplementary_Data
Click here for additional data file. (918.3KB, docx)

Acknowledgements

The authors thank Chao Ning, Yabing Hu, and Mingming Su for many helpful discussions.

Contributor Information

Xiaojing Wu, Basic Medical School, Anhui Medical University, Hefei 230022, China; National Center for Protein Sciences (Beijing), Beijing Proteome Research Center, Beijing Institute of Lifeomics, Beijing 102206, China.

Mingfei Han, National Center for Protein Sciences (Beijing), Beijing Proteome Research Center, Beijing Institute of Lifeomics, Beijing 102206, China.

Xinyu Song, Center for Artificial Intelligence in Medicine, Medical Innovation Research Division of Chinese, PLA General Hospital, Beijing 100853, China.

Song He, Department of Bioinformatics, Institute of Health Service and Transfusion Medicine, Beijing 100850, China.

Xiaochen Bo, Department of Bioinformatics, Institute of Health Service and Transfusion Medicine, Beijing 100850, China.

Yunping Zhu, Basic Medical School, Anhui Medical University, Hefei 230022, China; National Center for Protein Sciences (Beijing), Beijing Proteome Research Center, Beijing Institute of Lifeomics, Beijing 102206, China.

Supplementary data

Supplementary data are available at Bioinformatics online.

Conflict of interest

None declared.

Funding

This work was supported by the National Key Research and Development Program of China [2021YFA1301603] and National Natural Science Foundation of China [62103436].

Data availability

The COMMO server is designed to provide users with intuitive analysis of publicly available datasets of The Cancer Genome Atlas (TCGA), the International Cancer Genome Consortium (ICGC), and Therapeutically Applicable Research To Generate Effective Treatments (TARGET, https://ocg.cancer.gov/programs/target), and to allow users to upload their own data for analysis.

References

  1. Bergmann S, Ihmels J, Barkai N.. Iterative signature algorithm for the analysis of large-scale gene expression data. Phys Rev E Stat Nonlin Soft Matter Phys 2003;67:031902. [DOI] [PubMed] [Google Scholar]
  2. Brière G, Darbo É, Thébault P et al. Consensus clustering applied to multi-omics disease subtyping. BMC Bioinformatics 2021;22:361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bu F, Zhao Y, Zhao Y. et al. Distinct tumor microenvironment landscapes of rectal cancer for prognosis and prediction of immunotherapy response. Cell Oncol (Dordr) 2022;45:1363–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cabassi A, Kirk PDW.. Multiple kernel learning for integrative consensus clustering of omic datasets. Bioinformatics 2020;36:4789–96. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Fu L, Medico E.. FLAME, a novel fuzzy clustering method for the analysis of DNA microarray data. BMC Bioinformatics 2007;8:3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. He S, Xu W, He H. et al. ICM: a web server for integrated clustering of multi-dimensional biomedical data. Nucleic Acids Research 2016;44(W1):W154–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. He S, Song X, Yang X. et al. COMSUC: a web server for the identification of consensus molecular subtypes of cancer based on multiple methods and multi-omics data. PLoS Comput Biol 2021;17:e1008769. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Hoadley KA, Yau C, Wolf DM. et al. Multiplatform analysis of 12 cancer types reveals molecular classification within and across tissues of origin. Cell 2014;158:929–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Huynh-Thu VA, Irrthum A, Wehenkel L. et al. Inferring regulatory networks from expression data using tree-based methods. PLoS One 2010;5:e12776. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Hyvärinen A. Independent component analysis: recent advances. Philos Trans A Math Phys Eng Sci 2013;371:20110534. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Jojic V, Shay T, Sylvia K. et al. ; Immunological Genome Project Consortium. Identification of transcriptional regulators in the mouse immune system. Nat Immunol 2013;14:633–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Kandoth C, Schultz N; Cancer Genome Atlas Research Network. et al. Integrated genomic characterization of endometrial carcinoma. Nature 2013;497:67–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Kluger Y, Basri R, Chang JT. et al. Spectral biclustering of microarray data: coclustering genes and conditions. Genome Res 2003;13:703–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Li J, Miao B, Wang S. et al. Hiplot: a comprehensive and easy-to-use web service for boosting publication-ready biomedical data visualization. Brief Bioinform 2022;23:bbac261. [DOI] [PubMed] [Google Scholar]
  15. Liefeld T, Huang E, Wenzel AT. et al. NMFClustering: accessible NMF-based clustering utilizing GPU acceleration. bioRxiv 2023, preprint: not peer reviewed. [DOI] [PMC free article] [PubMed]
  16. Liu X, Xu K, Tao X. et al. ExpressVis: a biologist-oriented interactive web server for exploring multi-omics data. Nucleic Acids Res 2022a;50(W1):W312–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Liu Z, Li Y, Yao L. et al. Agglomerative neural networks for multiview clustering. IEEE Trans Neural Netw Learn Syst 2022b;33:2842–52. [DOI] [PubMed] [Google Scholar]
  18. Qin W, Gan F, Jike Y. et al. Consensus clustering and survival-related genes of cuproptosis in cutaneous melanoma. Mediators Inflamm 2023;2023:3615688. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Saelens W, Cannoodt R, Saeys Y.. A comprehensive evaluation of module detection methods for gene expression data. Nat Commun 2018;9:1090. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. The Cancer Genome Atlas Network. Comprehensive molecular portraits of human breast tumours. Nature 2012;490:61–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Timmerman ME, Ceulemans E, De Roover K. et al. Subspace K-means clustering. Behav Res Methods 2013;45:1011–23. [DOI] [PubMed] [Google Scholar]
  22. Wang J, Delabie J, Aasheim H. et al. Clustering of the SOM easily reveals distinct gene expression patterns: results of a reanalysis of lymphoma study. BMC Bioinformatics 2002;3:36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Wang W, Tan H, Sun M. et al. Independent component analysis based gene co-expression network inference (ICAnet) to decipher functional modules for better single-cell clustering and batch integration. Nucleic Acids Res 2021;49:e54. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Yu G, Wang LG, Han Y. et al. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics 2012;16:284–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Zhang X, Liu H, Wu XM. et al. Spectral embedding network for attributed graph clustering. Neural Netw 2021;142:388–96. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

btad708_Supplementary_Data
Click here for additional data file. (918.3KB, docx)

Data Availability Statement

The COMMO server is designed to provide users with intuitive analysis of publicly available datasets of The Cancer Genome Atlas (TCGA), the International Cancer Genome Consortium (ICGC), and Therapeutically Applicable Research To Generate Effective Treatments (TARGET, https://ocg.cancer.gov/programs/target), and to allow users to upload their own data for analysis.

Articles from Bioinformatics are provided here courtesy of Oxford University Press

RESOURCES