Consensus clustering applied to multi-omics disease subtyping

Abstract

Background: Facing the diversity of omics data and the difficulty of selecting one result over all those produced by several methods, consensus strategies have the potential to reconcile multiple inputs and to produce robust results.

Results: Here, we introduce ClustOmics, a generic consensus clustering tool that we use in the context of cancer subtyping. ClustOmics relies on a non-relational graph database, which allows for the simultaneous integration of both multiple omics data and results from various clustering methods. This new tool conciliates input clusterings, regardless of their origin, their number, their size or their shape. ClustOmics implements an intuitive and flexible strategy, based upon the idea of evidence accumulation clustering. ClustOmics computes co-occurrences of pairs of samples in input clusters and uses this score as a similarity measure to reorganize data into consensus clusters.

Conclusion: We applied ClustOmics to multi-omics disease subtyping on real TCGA cancer data from ten different cancer types. We showed that ClustOmics is robust to heterogeneous qualities of input partitions, smoothing and reconciling preliminary predictions into high-quality consensus clusters, both from a computational and a biological point of view. The comparison to a state-of-the-art consensus-based integration tool, COCA, further corroborated this statement. However, the main interest of ClustOmics is not to compete with other tools, but rather to make profit from their various predictions when no gold-standard metric is available to assess their significance.

Availability: The ClustOmics source code, released under MIT license, and the results obtained on TCGA cancer data are available on GitHub: https://github.com/galadrielbriere/ClustOmics .

Keywords: Consensus clustering; Data integration; Disease subtyping; Multi-omic data.

Fig. 1

Two integration scenarios: multi-to-multi consensus clustering and single-to-multi consensus clustering. Arrows are dashed according to the omics considered by each input clustering method

Fig. 2

Overview of survival and clinical label enrichment results for the ten cancer types analyzed. The x-axis represents the number of significant survival p values ($<0.01$) found for each clustering, over all the ten cancer types. The y-axis represents the total number of significantly enriched clinical labels (p values $<0.01$), all cancer types included. In total, 79 enrichment p values were computed from 32 distinct clinical labels

Fig. 3

Survival analysis results for ClustOmics and COCA multi-to-multi consensus clustering and for each input multi-omics clustering. The horizontal dashed line indicates the threshold for significantly different survival rate (p value $\le 0.01$). Boxplots were computed considering input clusterings only

Fig. 4

Adjusted Rand index of input clusterings relative to ClustOmics and COCA StoM consensus multi-omics clusterings. Each point corresponds to one clustering and is colored according to the omics type used. Each omics dataset was clustered using five different clustering tools (PINS, NEMO, SNF, rMKL, K-means), and therefore, it is represented by five input clusterings. ClustOmics and COCA respective consensus clustering similarity is displayed with a black square and a brown diamond

Fig. 5

Survival analysis results for ClustOmics and COCA single-to-multi All consensus clustering and for each input multi-omics clustering. The horizontal dashed line indicates the threshold for significantly different survival rate (p value $\le 0.01$). Boxplots were computed considering only input clusterings

Fig. 6

BIC consensus clustering with patients colored according to the PAM50 prediction. Annotated screenshot from the Neo4j browser for graph visualization

Fig. 7

An overview of the strategy implemented in ClustOmics

Fig. 8

An integration graph filtered with increasing threshold values: 1, 3, and 5 (the maximum number of supports for an integration edge being 5 in this example). Screenshots from the Neo4j browser for graph visualization