Open-access ‘knowledgebase’ focuses on clinical relevance of cancer gene mutations

Cancer researchers at Washington University School of Medicine in St. Louis have developed an online cancer database designed to streamline the sharing of tumor-related genetic information.

The open-access database — known as CIViC, which stands for Clinical Interpretation of Variants in Cancer — launched last year. It is accessible at civic.genome.wustl.edu.

It includes information about nearly 300 genes and 750 variants for 171 cancer types.

Although submissions are accepted, two independent contributors — one of whom must be an expert editor — must agree the information is worthy of inclusion.

Obi L. Griffith, PhD, and Malachi Griffith, PhD — identical twins who both serve as assistant professors at Washington University in St. Louis School of Medicine and assistant directors of the university’s McDonnell Genome Institute — co-lead the project.

HemOnc Today spoke with Obi Griffith about what the database offers, and its potential benefits for researchers and clinicians.

Question: How did the idea for CIViC come about , and why do you think it is necessary?

Answer: The idea for the CIViC ‘knowledgebase’ came out of our work with our genomics tumor board. We found that, as we were sequencing individual tumor cases, most of our time and effort was spent at the interpretation step: trying to take tens of thousands of genetic alterations that we identified by sequencing and determine which, if any, had clinical relevance in terms of predicting molecular subtype, predicting outcome or suggesting treatment options. This requires both genomics expertise to understand the nature of the molecular alteration, and an understanding of cancer biology and relevant clinical studies. It typically involves intensive manual labor, reading the biomedical literature and producing short synopses of the current relevance of any given variant. We found we were sometimes repeating this effort and losing the product of it by not storing the results systematically. We also knew that many others in the field were having the same difficulties. The idea of a common knowledgebase that allows external contributions solved a problem for ourselves while also providing a solution to a common problem for the wider community.

#
Obi L. Griffith, PhD, (left) and Malachi Griffith, PhD, (right)

Q: What distinguishes CIViC from other cancer databases?

A: Many other databases are more like variant warehouses compared with CIViC. We’re not trying to create a comprehensive list of all the cancer variants that have ever been seen. We are focused on the clinical relevance of these variants. Only a tiny fraction of all variants observed in sequencing studies are shown to have any specific clinical relevance. Partly that’s because there are many “passenger mutations,” where a gene has hundreds — sometimes thousands — of random mutations, but only one or a few of them cause cells to grow out of control and proliferate. Identifying relevant mutations requires extensive functional studies, lab experiments and eventually clinical trials that investigate the sensitivity of one therapy vs. another in the context of the mutation. It has become very time-consuming for the typical clinician or genome scientist to Google or search PubMed and read about a mutation, then try to remind themselves what — if any — clinical relevance it has. What CIViC is really focused on is that problem: carefully documenting, in one location, the evidence for clinical relevance of as many mutations as possible.

Q: Why did you decide to make it open access?

A: An open-source, open-access approach maximizes the benefit of CIViC for the research community. We present no legal or commercial barriers to using the knowledge accumulated in CIViC. This allows both academic and commercial groups to rapidly incorporate CIViC knowledge into their studies, workflows or products. It also allows us to benefit from the contributions of other experts around the world. They can contribute, knowing that their work will remain available for their own use and the use of others.

Q: How can someone go about contributing to CIViC? How many users have contributed?

A: Anyone can create an account using their existing Google, Github or ORCid account and start as a “CIViC curator.” This means that the user can comment on existing evidence, variant or gene-level interpretations, suggest new sources or publications for curation, and suggest specific changes or additions to the knowledgebase. All of this is done by simply navigating and interacting with the website. A CIViC editor must then review and approve those suggestions or additions. As of early March, there were 64 active contributors to the knowledgebase, with many more users utilizing the data without contributing.

Q: Can you describe the curation/moderation process?

A: The curation and moderation process involves logging into the CIViC interface and completing “curation activities.” For new curators, these can be as small as suggesting a new publication or making a minor edit, or more time-consuming, such as submitting evidence for a new variant interpretation, entering variant coordinates, or writing gene or variant-level summaries. For editors, most of their time is spent monitoring new submissions or revisions, reviewing those suggestions, commenting, and ultimately approving or rejecting them.

Q: How can CIViC benefit oncologists trying to understand and interpret data about tumor genetics?

A: If an oncologist is presented with a patient with a specific molecular alteration, they can simply search the website for that gene or variant and review the summaries and evidence for the clinical relevance of that gene or variant in a specific tumor type. This will hopefully present a shortcut to relevant biomedical research results compared with simply searching PubMed, the current paradigm for many. A main advantage CIViC has over other resources is that if a user identifies an error or knowledge gap, there is a straightforward procedure to flag the issue, or even propose a correction or addition. Increasingly, CIViC interpretations are being incorporated automatically into reports generated from clinical sequencing workflows. We hope to develop additional tools that provide variant-interpretation mapping, knowledge-driven gene panel sequencing and automated report generation.

Q: How do you envision CIViC evolving?

A: We are working to improve the interface in terms of trying to increase the ease of use and engagement with expert curators. We’re also working hard to coordinate efforts across databases to consolidate information. In addition, we have some large imports of data underway from other groups that have curated the information but don’t have a system of their own for making those variants available to the public. – by Kyle Doherty

For more information:

Obi L. Griffith , PhD, can be reached at McDonnell Genome Institute, Washington University, Campus Box 8501, 4444 Forest Park Ave., St. Louis, MO 63108; email: obigriffith@wustl.edu.

Cancer researchers at Washington University School of Medicine in St. Louis have developed an online cancer database designed to streamline the sharing of tumor-related genetic information.

The open-access database — known as CIViC, which stands for Clinical Interpretation of Variants in Cancer — launched last year. It is accessible at civic.genome.wustl.edu.

It includes information about nearly 300 genes and 750 variants for 171 cancer types.

Although submissions are accepted, two independent contributors — one of whom must be an expert editor — must agree the information is worthy of inclusion.

Obi L. Griffith, PhD, and Malachi Griffith, PhD — identical twins who both serve as assistant professors at Washington University in St. Louis School of Medicine and assistant directors of the university’s McDonnell Genome Institute — co-lead the project.

HemOnc Today spoke with Obi Griffith about what the database offers, and its potential benefits for researchers and clinicians.

Question: How did the idea for CIViC come about , and why do you think it is necessary?

Answer: The idea for the CIViC ‘knowledgebase’ came out of our work with our genomics tumor board. We found that, as we were sequencing individual tumor cases, most of our time and effort was spent at the interpretation step: trying to take tens of thousands of genetic alterations that we identified by sequencing and determine which, if any, had clinical relevance in terms of predicting molecular subtype, predicting outcome or suggesting treatment options. This requires both genomics expertise to understand the nature of the molecular alteration, and an understanding of cancer biology and relevant clinical studies. It typically involves intensive manual labor, reading the biomedical literature and producing short synopses of the current relevance of any given variant. We found we were sometimes repeating this effort and losing the product of it by not storing the results systematically. We also knew that many others in the field were having the same difficulties. The idea of a common knowledgebase that allows external contributions solved a problem for ourselves while also providing a solution to a common problem for the wider community.

#
Obi L. Griffith, PhD, (left) and Malachi Griffith, PhD, (right)

Q: What distinguishes CIViC from other cancer databases?

A: Many other databases are more like variant warehouses compared with CIViC. We’re not trying to create a comprehensive list of all the cancer variants that have ever been seen. We are focused on the clinical relevance of these variants. Only a tiny fraction of all variants observed in sequencing studies are shown to have any specific clinical relevance. Partly that’s because there are many “passenger mutations,” where a gene has hundreds — sometimes thousands — of random mutations, but only one or a few of them cause cells to grow out of control and proliferate. Identifying relevant mutations requires extensive functional studies, lab experiments and eventually clinical trials that investigate the sensitivity of one therapy vs. another in the context of the mutation. It has become very time-consuming for the typical clinician or genome scientist to Google or search PubMed and read about a mutation, then try to remind themselves what — if any — clinical relevance it has. What CIViC is really focused on is that problem: carefully documenting, in one location, the evidence for clinical relevance of as many mutations as possible.

PAGE BREAK

Q: Why did you decide to make it open access?

A: An open-source, open-access approach maximizes the benefit of CIViC for the research community. We present no legal or commercial barriers to using the knowledge accumulated in CIViC. This allows both academic and commercial groups to rapidly incorporate CIViC knowledge into their studies, workflows or products. It also allows us to benefit from the contributions of other experts around the world. They can contribute, knowing that their work will remain available for their own use and the use of others.

Q: How can someone go about contributing to CIViC? How many users have contributed?

A: Anyone can create an account using their existing Google, Github or ORCid account and start as a “CIViC curator.” This means that the user can comment on existing evidence, variant or gene-level interpretations, suggest new sources or publications for curation, and suggest specific changes or additions to the knowledgebase. All of this is done by simply navigating and interacting with the website. A CIViC editor must then review and approve those suggestions or additions. As of early March, there were 64 active contributors to the knowledgebase, with many more users utilizing the data without contributing.

Q: Can you describe the curation/moderation process?

A: The curation and moderation process involves logging into the CIViC interface and completing “curation activities.” For new curators, these can be as small as suggesting a new publication or making a minor edit, or more time-consuming, such as submitting evidence for a new variant interpretation, entering variant coordinates, or writing gene or variant-level summaries. For editors, most of their time is spent monitoring new submissions or revisions, reviewing those suggestions, commenting, and ultimately approving or rejecting them.

Q: How can CIViC benefit oncologists trying to understand and interpret data about tumor genetics?

A: If an oncologist is presented with a patient with a specific molecular alteration, they can simply search the website for that gene or variant and review the summaries and evidence for the clinical relevance of that gene or variant in a specific tumor type. This will hopefully present a shortcut to relevant biomedical research results compared with simply searching PubMed, the current paradigm for many. A main advantage CIViC has over other resources is that if a user identifies an error or knowledge gap, there is a straightforward procedure to flag the issue, or even propose a correction or addition. Increasingly, CIViC interpretations are being incorporated automatically into reports generated from clinical sequencing workflows. We hope to develop additional tools that provide variant-interpretation mapping, knowledge-driven gene panel sequencing and automated report generation.

PAGE BREAK

Q: How do you envision CIViC evolving?

A: We are working to improve the interface in terms of trying to increase the ease of use and engagement with expert curators. We’re also working hard to coordinate efforts across databases to consolidate information. In addition, we have some large imports of data underway from other groups that have curated the information but don’t have a system of their own for making those variants available to the public. – by Kyle Doherty

For more information:

Obi L. Griffith , PhD, can be reached at McDonnell Genome Institute, Washington University, Campus Box 8501, 4444 Forest Park Ave., St. Louis, MO 63108; email: obigriffith@wustl.edu.