Prioritizing Curation Effort¶
New CIViC curators, commonly ask where they should focus their efforts and where can they find evidence of the significance of cancer variants. There are many approaches and relevant resources that may help to identify and prioritize such evidence.
Remember that the focus of CIViC is on the clinical relevance of cancer variants, which CIViC describes using Molecular Profiles (MPs) - collections of one or more variants, associated to one or more genes. Before expending the effort to propose an addition to CIViC, ask yourself would a clinician potentially find this information useful in understanding and treating a patient’s cancer? Could an oncologist use this evidence to better understand likely response to therapy (Predictive evidence), or outcome (Prognostic) for their patient? Would a pathologist or genetic laboratory director find knowledge of the Molecular Profile (variant) valuable in classifying (Diagnostic) the tumor into a subtype? Would a medical geneticist or genetic counselor be interested in the causative (Predisposing) significance of this evidence? Does the variant occur in a gene of known significance to cancer (in this case a single variant, simple Molecular Profile)? Will determining the Oncogenic or Functional potential of the Molecular Profile help to interpret its potential clinical relevance?
In addressing these questions, try to think about the distinction between the relevance of a Molecular Profile (MP) to cancer biology and its relevance in a clinical setting. An MP may have great and diverse relevance to the biology of a cancer cell but have limited or no clinical applicability. For example, TP53 mutations are critical in many cancers and hundreds (if not thousands) of papers have been written about their complex roles in cancer biology. However, the scenarios in which TP53 mutations are clinically relevant are much, much narrower. An MP may NOT be clinically relevant despite being characterized as functional (gain or loss of function), a ‘driver’, ‘recurrent’, etc. In some, perhaps most cases, the clinical relevance of these MPs may simply not be established yet. However, CIViC is about the evidence that establishes their clinical relevance. By contrast, in some cases, the biological relevance may be poorly understood while clinical utlity is established. Such evidence does belong in CIViC. A mechanistic understanding is highly desirable, but not strictly required.
The above description is an oversimplification. The concept of clinical utility varies by Evidence type: Predictive, Prognostic, Diagnostic, Predisposing, Oncogenic and Functional. The “ideal” clinical interpretation and definition of clinical utility are open to debate. We welcome this debate and one of the goals of CIViC is to enable and capture it. If you believe some evidence is relevant to CIViC but have some doubts, please submit it so that the community can discuss with you.
The following list is not exhaustive but provides many examples of approaches to identify high quality evidence. If you know of a useful resource that is not listed below, please let us know about it. NOTE: some of these resources are open access, others are not. When entering evidence into CIViC, never copy content or ideas from another resource. Your contributions to CIViC should be based on published evidence, but in your own words.
Example sources of CIViC evidence and high priority variants
Published results from clinical trials involving cancers with specific variants (e.g. HER2 +ve breast cancer)
A gene, variant, simple or complex Molecular Profile, or paper, that you are an expert in. For example, this might be work from your own research/practice.
Public discussions on cases submitted to the ASCO Molecular Oncology Tumor Board
The CIViC publication queue, a place were CIViC curators add and discuss papers thought to contain valuable evidence.
We created a ranked list of relevant genes, based on a comprehensive survey of genes that are targeted by dozens of assays in clinical use.
We also created a ranked list of relevant publications by summarizing overlap between the publications used in CIViC and other companion resources.
Our colleagues at the BC Cancer Agency and Glasgow University have developed a natural language processing approach and resulting database of automatically mined CIViC relevant publications called: CIViC-mine.
Companion resources of CIViC participating in the GA4GH Variant Interpretation for Cancer Consortium (VICC) or others such as: PMKB, OncoKB, MyCancerGenome, CanDL, BaseSpace KN, Cancer Genome Interpreter, COSMIC, PCT, PharmGKB. A detailed comparison of these resources can be found in the CIViC Related Resources Table. While these resources can be used for inspiration, do not plagiarize/copy any content from these sources that might violate their copyrights.
Papers referenced by the Atlas of Genetics and Cytogenetics in Oncology and Haematology
Papers from certain topical journals. The most cited journals in CIViC are summarized on the CIViC Source Statistics page
Variants and related papers from the Sarcoma Initiative.
Variants from the LOVD project.
Example Curation Activities¶
Most of the contents of CIViC are curator generated and edited. Below are a few examples of areas where curators are needed, culminating in a list of bite sized curation tasks that should require less than 3 minutes of your time.
1. Adding Evidence
Evidence records are the heart of the CIViC resource. Each corresponds to evidence for a single clinical assertion about a molecular profile (single variant or combination of variants) in a single cancer type. Each evidence record is based on a single citable source (e.g. a peer-reviewed publication). The evidence record consists of a free-form executive summary describing the assertion and supporting evidence for it, and additional structured fields that describe the evidence (e.g. evidence type, relevant therapy, etc.). To learn more about the elements of an evidence record refer to the Evidence help docs. For suggestions on where to find sources for evidence records, refer to the Source Ideas tab of this section. To add a new evidence record, you must login and hit the “ADD” button at the top of any page throughout the site.
2. Adding or improving Molecular Profile, Variant Group, and Gene Descriptions
Once sufficient evidence accumulates for a molecular profile (single variant or combination of variants), variant group or gene entity, a summary description should be created. The summary should be an overview of the evidence records associated with the entity. The summary should focus on the most clinically actionable evidence and should summarize for each relevant gene what a clinician should be aware of for patients with a particular variant, or variants in a particular gene. Very brief background material may be included. Additional citations beyond those associated with the evidence records can be associated directly with the summary using the “Add Source” option in the edit form. To learn more about each major CIViC entity, refer to the Molecular Profiles, Variant Groups, and Genes sections of the help pages.
3. Adding or improving Assertions
An important final product of the CIViC curation process is the Assertion. Gene and Molecular Profile Summaries (described above) provide an overall summary of the clinical relevance of genes and molecular profiles as documented by the entire body of CIViC evidence. In contrast, Assertions provide a consensus of the significance (and supporting evidence) for a specific gene-molecular profile, in a specific disease context. The assertion should represent the current state of understanding in the field and be associated with the appropriate AMP tier or ACMG codes and assessment for the molecular profile. Once sufficient evidence has been documented, a new assertion can be submitted using the “ADD” button at the top of any page throughout the site. Reviewed and accepted assertions enter the queue for submission to ClinVar. Creating assertions is one of the most advanced curation tasks in CIViC.
4. Editing CIViC Content
CIViC content can be edited by clicking on the “Revise” button in the top right of any editable page. Gene, Molecular Profiles, Variant, Variant Group and Evidence entities can all be edited. These edits may be expansive major updates to incorporate new evidence, error corrections, improvements to readability and style, or minor grammar and typo fixes. All such edits are welcome.
5. Comment on CIViC Content
Throughout the website are “Comment” tabs where users can comment on the current contents of CIViC (specific Evidence, Molecular Profiles, Variants or Genes) or on Revisions. Curators are encouraged to be verbose in their comments on existing content. Critism, clarification, qualification, and questions are all appropriate. Comments from the authors of work being summarized or others with particular expertise in the area are especially desirable. When adding new evidence or summaries, comments may be used to describe the thought process of the curator. Small quotes (as allowed by the Fair Use doctrine) from source publications that support a submission may also be included (but please indicate these with quotes or use the block quote style).
6. Molecular Profile Description
The Molecular Profile (MP) page contains a Description that is written by curators, and should summarize the main ideas from the Evidence Items (EIDs) associated to the MP. The Description also allows for the citing of source publications used to incorporate other relevant information for the role of the MP in cancer.
7. Variant attributes
A Molecular Profile consists of combinations of one or more variants. Each variant also has its own page in CIViC. Variants have several structured values associated with variant records. These include:
Aliases. Alternative names (synonyms) for the variant. For many variants, researchers from different groups may refer to variants by different names. Multiple and varying abbreviations or identifiers exist for most variants. A variant alias is generally any name that might help CIViC users determine the various ways used to indicate the same variant.
HGVS expressions. CIViC supports and promotes variant identification using the Sequence Variant Nomenclature guidelines of the Human Genome Variation Society (HGVS), otherwise known as ‘HGVS strings’. Curators may add one or more valid HGVS values for each variant. These may be entered in protein (p.), cDNA (c.), or genomic (g.) format. A particular CIViC variant (e.g. BRAF V600E) may have multiple valid genomic alterations that could create it, each with a distinct genomic HGVS expression. Similarly, multiple cDNA HGVS strings may correspond to multiple transcript sequences, possibly from various transcript annotation databases (e.g. Ensembl, RefSeq, LRG, etc.) or alternative isoforms of a gene.
Coordinates. For each variant, the goal of CIViC initially is to determine unambiguous genomic coordinates for an example instance of the variant. For instance, if the paper refers to the variant as “V600E”, the curator determines for a particular build of the human genome, the corresponding chromosome, start position, end position, reference base and variant base. Refer to the Variants documentation on the left for more details.
8. Bite-size curation tasks
Only have a few minutes? Tackle one of the tasks below.
Identify a publication containing a variant with clinical relevance.
Visit PubMed to identify the publication’s PubMed ID.
Enter as much information as possible to help curators. This form only requires 2 elements: PubMed ID and a comment to direct curators as to why you believe this publication has clinically-relevant information about a variant.
Add a variant Alias.
Read a summary for your favorite gene/variant and comment on the contents.