Framework for quality assessment of whole genome cancer sequences

Bringing together cancer genomes from different projects increases power and allows the investigation of pan-cancer, molecular mechanisms. However, working with whole genomes sequenced over several years in different sequencing centres requires a framework to compare the quality of these sequences. We used the Pan-Cancer Analysis of Whole Genomes cohort as a test case to construct such a framework. This cohort contains whole cancer genomes of 2832 donors from 18 sequencing centres. We developed a non-redundant set of five quality control (QC) measurements to establish a star rating system. These QC measures reflect known differences in sequencing protocol and provide a guide to downstream analyses and allow for exclusion of samples of poor quality. We have found that this is an effective framework of quality measures. The implementation of the framework is available at: https://dockstore.org/containers/quay.io/jwerner_dkfz/pancanqc:1.2.2.

Working with cancer genomes from multiple projects can increase investigative power, but quality of sequences can vary. Here, the authors present a framework for comparing whole genome sequencing quality to help researchers guide downstream analyses and exclude poor quality samples.

Pan-cancer analysis of whole genomes

Cancer is driven by genetic change, and the advent of massively parallel sequencing has enabled systematic documentation of this variation at the whole-genome scale1-3. Here we report the integrative analysis of 2,658 whole-cancer genomes and their matching normal tissues across 38 tumour types from the Pan-Cancer Analysis of Whole Genomes (PCAWG) Consortium of the International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA). We describe the generation of the PCAWG resource, facilitated by international data sharing using compute clouds. On average, cancer genomes contained 4-5 driver mutations when combining coding and non-coding genomic elements; however, in around 5% of cases no drivers were identified, suggesting that cancer driver discovery is not yet complete. Chromothripsis, in which many clustered structural variants arise in a single catastrophic event, is frequently an early event in tumour evolution; in acral melanoma, for example, these events precede most somatic point mutations and affect several cancer-associated genes simultaneously. Cancers with abnormal telomere maintenance often originate from tissues with low replicative activity and show several mechanisms of preventing telomere attrition to critical levels. Common and rare germline variants affect patterns of somatic mutation, including point mutations, structural variants and somatic retrotransposition. A collection of papers from the PCAWG Consortium describes non-coding mutations that drive cancer beyond those in the TERT promoter4; identifies new signatures of mutational processes that cause base substitutions, small insertions and deletions and structural variation5,6; analyses timings and patterns of tumour evolution7; describes the diverse transcriptional consequences of somatic mutation on splicing, expression levels, fusion genes and promoter activity8,9; and evaluates a range of more-specialized features of cancer genomes8,10-18.

Recurrent somatic mutations reveal new insights into consequences of mutagenic processes in cancer

The sheer size of the human genome makes it improbable that identical somatic mutations at the exact same position are observed in multiple tumours solely by chance. The scarcity of cancer driver mutations also precludes positive selection as the sole explanation. Therefore, recurrent mutations may be highly informative of characteristics of mutational processes. To explore the potential, we use recurrence as a starting point to cluster >2,500 whole genomes of a pan-cancer cohort. We describe each genome with 13 recurrence-based and 29 general mutational features. Using principal component analysis we reduce the dimensionality and create independent features. We apply hierarchical clustering to the first 18 principal components followed by k-means clustering. We show that the resulting 16 clusters capture clinically relevant cancer phenotypes. High levels of recurrent substitutions separate the clusters that we link to UV-light exposure and deregulated activity of POLE from the one representing defective mismatch repair, which shows high levels of recurrent insertions/deletions. Recurrence of both mutation types characterizes cancer genomes with somatic hypermutation of immunoglobulin genes and the cluster of genomes exposed to gastric acid. Low levels of recurrence are observed for the cluster where tobacco-smoke exposure induces mutagenesis and the one linked to increased activity of cytidine deaminases. Notably, the majority of substitutions are recurrent in a single tumour type, while recurrent insertions/deletions point to shared processes between tumour types. Recurrence also reveals susceptible sequence motifs, including TT[C>A]TTT and AAC[T>G]T for the POLE and 'gastric-acid exposure' clusters, respectively. Moreover, we refine knowledge of mutagenesis, including increased C/G deletion levels in general for lung tumours and specifically in midsize homopolymer sequence contexts for microsatellite instable tumours. Our findings are an important step towards the development of a generic cancer diagnostic test for clinical practice based on whole-genome sequencing that could replace multiple diagnostics currently in use.

Consensus and conflict cards for metabolic pathway databases

Background

The metabolic network of H. sapiens and many other organisms is described in multiple pathway databases. The level of agreement between these descriptions, however, has proven to be low. We can use these different descriptions to our advantage by identifying conflicting information and combining their knowledge into a single, more accurate, and more complete description. This task is, however, far from trivial.

Results

We introduce the concept of Consensus and Conflict Cards (C₂Cards) to provide concise overviews of what the databases do or do not agree on. Each card is centered at a single gene, EC number or reaction. These three complementary perspectives make it possible to distinguish disagreements on the underlying biology of a metabolic process from differences that can be explained by different decisions on how and in what detail to represent knowledge. As a proof-of-concept, we implemented C₂Cards^Human, as a web application http://www.molgenis.org/c2cards, covering five human pathway databases.

Conclusions

C₂Cards can contribute to ongoing reconciliation efforts by simplifying the identification of consensus and conflicts between pathway databases and lowering the threshold for experts to contribute. Several case studies illustrate the potential of the C₂Cards in identifying disagreements on the underlying biology of a metabolic process. The overviews may also point out controversial biological knowledge that should be subject of further research. Finally, the examples provided emphasize the importance of manual curation and the need for a broad community involvement.

Knowledge representation in metabolic pathway databases

The accurate representation of all aspects of a metabolic network in a structured format, such that it can be used for a wide variety of computational analyses, is a challenge faced by a growing number of researchers. Analysis of five major metabolic pathway databases reveals that each database has made widely different choices to address this challenge, including how to deal with knowledge that is uncertain or missing. In concise overviews, we show how concepts such as compartments, enzymatic complexes and the direction of reactions are represented in each database. Importantly, also concepts which a database does not represent are described. Which aspects of the metabolic network need to be available in a structured format and to what detail differs per application. For example, for in silico phenotype prediction, a detailed representation of gene-protein-reaction relations and the compartmentalization of the network is essential. Our analysis also shows that current databases are still limited in capturing all details of the biology of the metabolic network, further illustrated with a detailed analysis of three metabolic processes. Finally, we conclude that the conceptual differences between the databases, which make knowledge exchange and integration a challenge, have not been resolved, so far, by the exchange formats in which knowledge representation is standardized.

Critical assessment of human metabolic pathway databases: a stepping stone for future integration

Background

Multiple pathway databases are available that describe the human metabolic network and have proven their usefulness in many applications, ranging from the analysis and interpretation of high-throughput data to their use as a reference repository. However, so far the various human metabolic networks described by these databases have not been systematically compared and contrasted, nor has the extent to which they differ been quantified. For a researcher using these databases for particular analyses of human metabolism, it is crucial to know the extent of the differences in content and their underlying causes. Moreover, the outcomes of such a comparison are important for ongoing integration efforts.

Results

We compared the genes, EC numbers and reactions of five frequently used human metabolic pathway databases. The overlap is surprisingly low, especially on reaction level, where the databases agree on 3% of the 6968 reactions they have combined. Even for the well-established tricarboxylic acid cycle the databases agree on only 5 out of the 30 reactions in total. We identified the main causes for the lack of overlap. Importantly, the databases are partly complementary. Other explanations include the number of steps a conversion is described in and the number of possible alternative substrates listed. Missing metabolite identifiers and ambiguous names for metabolites also affect the comparison.

Conclusions

Our results show that each of the five networks compared provides us with a valuable piece of the puzzle of the complete reconstruction of the human metabolic network. To enable integration of the networks, next to a need for standardizing the metabolite names and identifiers, the conceptual differences between the databases should be resolved. Considerable manual intervention is required to reach the ultimate goal of a unified and biologically accurate model for studying the systems biology of human metabolism. Our comparison provides a stepping stone for such an endeavor.