Establishing a consensus for the hallmarks of cancer based on gene ontology and pathway annotations

Abstract

Background: The hallmarks of cancer provide a highly cited and well-used conceptual framework for describing the processes involved in cancer cell development and tumourigenesis. However, methods for translating these high-level concepts into data-level associations between hallmarks and genes (for high throughput analysis), vary widely between studies. The examination of different strategies to associate and map cancer hallmarks reveals significant differences, but also consensus.

Results: Here we present the results of a comparative analysis of cancer hallmark mapping strategies, based on Gene Ontology and biological pathway annotation, from different studies. By analysing the semantic similarity between annotations, and the resulting gene set overlap, we identify emerging consensus knowledge. In addition, we analyse the differences between hallmark and gene set associations using Weighted Gene Co-expression Network Analysis and enrichment analysis.

Conclusions: Reaching a community-wide consensus on how to identify cancer hallmark activity from research data would enable more systematic data integration and comparison between studies. These results highlight the current state of the consensus and offer a starting point for further convergence. In addition, we show how a lack of consensus can lead to large differences in the biological interpretation of downstream analyses and discuss the challenges of annotating changing and accumulating biological data, using intermediate knowledge resources that are also changing over time.

Keywords: Co-expression network; Gene ontolog; Semantic similarity; The hallmarks of cancer.

Fig. 1

a Frequency of selection of GO terms from different schemes. b Frequency of selection of GO terms for individual cancer hallmarks. The x-axis represents the number of GO terms. The Y-axis represents individual cancer hallmarks. Bars coloured in different color represent how frequently it was selected by mapping methods to annotate this cancer hallmark

Fig. 2

Hallmark Gene Set Comparison. The upset plot shows the number of genes in each hallmark gene set and their intersections. The orange and blue lines represent genes shared by GO mapping schemes and all mapping schemes respectively

Fig. 3

Pairwise comparison of groups of prognostic-hallmark genes identified in the same cancers between mapping schemes. Each block represents the overlap of a group of prognostic-hallmark genes shared by multiple cancer types with different mapping schemes. The X-axis shows pairwise comparison of mapping schemes. The color of each block represents the similarity scores of the same group of genes using the Jaccard index. Red represents a higher score and Blue represents a lower score

Fig. 4

a Network analysis of prognostic hallmark gene expression in breast cancer with GO1 schemes identifies distinct modules of closely interconnected genes. b PPI network of hub genes of Module GO3_2, GO2_3 and GO1_2

Fig. 5

a Semantic similarity scores between modules. Red indicates high similarity and blue represents low similarity. b The top 10 significant GO terms (P< 0.05) in GO1_2, GO2_3 and GO3_2 modules

Fig. 6

A comparison of the GO Biological Process topology of GO terms selected by GO1 (2016) and GO2 (2012), constructed from all selected GO terms and their first neighbours. Crimson nodes represent GO terms that existed in both time points but were only selected by GO2, while light red nodes represent GO terms which were obsoleted in 2016. Similarly, dark blue nodes represent GO terms that existed in both 2012 and 2016 but were only selected by GO1 and light blue nodes represent GO terms had not been created in 2012

Fig. 7

Visualization of the consensus GO terms for defining the cancer hallmark ‘Activating Invasion and Metastasis’. Nodes colored in red were selected by 3 schemes and nodes coloured in yellow represent GO terms selected by 4 methods