ClinVar Miner: Demonstrating utility of a Web-based tool for viewing and filtering ClinVar data

Abstract

ClinVar Miner is a Web-based suite that utilizes the data held in the National Center for Biotechnology Information's ClinVar archive. The goal is to render the data more accessible to processes pertaining to conflict resolution of variant interpretation as well as tracking details of data submission and data management for detailed variant curation. Here, we establish the use of these tools to address three separate use cases and to perform analyses across submissions. We demonstrate that the ClinVar Miner tools are an effective means to browse and consolidate data for variant submitters, curation groups, and general oversight. These tools are also relevant to the variant interpretation community in general.

Keywords: ClinVar; clinical domain working group; expert panel; variant archive; variant curation; variant interpretation.

Figure 1

ClinVar review status designations for submissions and overall variant clinical significance. ClinVar reports the level of review supporting the assertion of clinical significance for the individual submission and for the aggregated variation.

Figure 2

A display of the submitter and significance breakdown for the APC gene within the Variants by gene tool. The screen shot displayed has been filtered to include only submitters with submissions with assertion criteria provided but all submitters can be listed if these filters are not applied. This tool can help identify labs to participate in clinical domain working groups and expert panels.

Figure 3

A workflow showing the Variants in conflict by submitter tool for identifying conflicting interpretations between two individual submitters. The filters at the top of the page apply to all submissions included in subsequent displays. The Minimum conflict level filter allows the user to display For example, setting the review status as ‘criteria provided’ and collection method as ‘clinical testing’ of the submitter of interest and the comparator and setting the minimum conflict level at ‘benign or likely benign vs uncertain conflict’ and clicking the Apply filters button (A) will exclude all submissions that are ‘no criteria provided’ and are not ‘clinical testing’ and will consider as conflicts only LB/B vs VUS conflicts, category conflicts, and clinically significant conflicts. When one or more filters are applied, all calculated conflicts and variants reflect only conflicts or submissions that pass the filter(s). Selecting a laboratory from the list (B) will open a new page displaying a summary of conflicts from that submitter as well as a list of conflicting submitters. Selecting one of those submitters (C) opens a new page displaying only conflicts between the two submitters. Selecting this cell in the table (D) displays a list of variants that the LMM submitted as likely benign and GeneDx submitted as pathogenic.

Figure 4

Views of the high-level trends in ClinVar Miner. A) Submissions over time broken down by collection method. B) A choropleth showing number of submissions by country. C) The cumulative number of significance terms submitted to the database over time.

Figure 5

Concordance and discordance compared to clinical testing submissions. Interpretations from clinical testing submissions, as the largest reference set, were compared to interpretations from each collection method. This data was generated with the Variants in conflict by submitter tool. The “Other” category includes the collection methods case-control, in vivo, in vitro, reference population, provider interpretation, and phenotyping only. Each variant with more than one conflict is counted as its highest conflict level, with clinically significant higher than category which is higher than B/LB vs VUS. Confidence conflicts were not considered conflicts in this analysis.

Figure 6

The top 10 genes with the most variants submitted to ClinVar. The Variants by gene tool was used to generate this data. For each of these genes the number and types of conflict are also shown. Only submissions normalized to the five standard significance terms were included. The Variants in conflict by gene tool was used to provide the number of conflicting submissions for each gene. The ATM gene also contains variants that affect both ATM and C11orf65, an overlapping non-coding gene.

Figure 7

Frequency of clinically significant conflicting interpretations of variants per gene. The gray region represents genes with only one submitter within which conflicts could not occur. Variants were limited to those affecting only one gene, thereby excluding larger copy number variants.