HVP5: 5th Biennial Meeting of the Human Variome Project Consortium
HVP's fifth biennial meeting was held at UNESCO in Paris on 20 to 22 May 2014.
Several other meetings, workshops and training courses were also conducted in conjunction with HVP5. These included:
- Gene/Disease Specific Database Curation Training Course
- HVP Country Node Establishment workshop: new directions
- Breast Cancer Interest Group (invitation only; view minutes)
- Oesophageal Cancer Meeting (invitation only)
- HVP International Scientific Advisory Committee meeting (view minutes)
- HVP Gene and Disease Specific Database Advisory Council meeting (view minutes)
- HVP International Confederation of Countries Advisory Council meeting (view minutes)
- HVPIL Board meeting
Partnerships, networking and collaboration
HVP5 marked the mid-point in Project Roadmap 2012–2016 and as such provided an important opportunity to both review past achievements and to refine the strategic goals of the project as it moves into the second half of the current strategic plan. Several announcements were made regarding the restructuring of the governance of the Human Variome Project that go to strengthening its operations and global networking and included:
- Increased regional representation on the Board of Human Variome Project International Limited
- New Scientific Directors and Professor Richard Cotton moving from Scientific Director to a new role of Founding Patron
The Human Variome Project recognizes the importance of working with key partners to achieve its goals. Other organizations operating in similar goals and mandates including those that emphasize the importance of quality basic sciences research in the area of human genetics and genomics, the practical application of this knowledge to developing effective and efficient services for patients, together with a strong emphasis on open and transparent access to variant data and information in all parts of the world can often contribute. Collaborations with UNESCO, the UN organization with the mandate for science and education are on track and will be taken to a higher level with further recognition of the Human Variome Project as an affiliated NGO of UNESCO to happen later in 2014.
Discussions with the World Health Organisation are likely to give rise to a formal agreement to work together, again later in 2014. This will assist the Human Variome Project to achieve its goals in the area of health and disease management, particularly in low and middle income countries.
Ties with HUGO have been further cemented by agreeing to host sessions at each other's international meetings and to embark on a joint project, together with HGVS aimed at developing a sustainable process for agreeing and promoting the international nomenclature for describing sequence variants. Other interested organizations will be invited to contribute to this work.
Discussions are also underway with the Global Alliance for Genomics and Health (GA4GH) about areas of common interest. Many Human Variome Project Consortium members are already involved in GA4GH working groups and committees and HVP's Breast Cancer Interest Group (view meeting minutes) looks likely to join with a sub-group of GA4GH to better co-ordinate and harmonise global efforts related to a specific group of variant genes.
The Human Variome Project International Scientific Advisory Committee has a newly elected committee and has formalized its operations to include regular monthly meetings and will focus on five key priorities: improving quality assurance of data collection and curation, developing capacity through education, emphasizing fiscal sustainability for variant data and information collection storage and sharing, supporting global partnerships, promoting ethical practice
The definition of HVP Country Nodes, what they do, their establishment and how they operate underwent a major review during 2013, with a new process being introduced on 1 January 2014. This has proved positive and has captured the growing interest in the coordinating the generation of national data that is occurring in many parts of the world. The number of HVP Country Nodes has doubled in the past 18 months. The ICCAC will focus its efforts on defining minimum content for national variant databases and new Working Groups will be established to address data quality, phenotype, and ethics.
Current developments in Africa
A major theme of the meeting was to focus on human genomics and genetics activities in the African region and to hear first-hand of the range and extent of activities currently being undertaken in this regional of the world. This is a largely under-studied community. A representative group of highly experienced and distinguished practitioners currently working in selected countries - Nigeria, Botswana, Uganda, South Africa, Cameroon, Mali and Ghana - were invited to present their current work. These researchers are increasingly connected to collaborators across the region, as well as to those in other parts of the world. While changes are being made, the challenges faced by many professionals working in continental Africa cover a range of issues, some of which are not so obvious to those who are not familiar with the region. For example:
- Weak legislative frameworks to help clarify ownership of data and information, determine appropriate policies to govern who has access to various types of data and under what conditions, including the movement of samples between countries
- Weak legislative and regulatory frameworks to protect the rights of patients and to deal pro-actively with all the relevant ethical, legal, social and cultural issues in diverse settings
- Need to strengthen information management systems in general and strategically plan for future needs
- Dealing with increased pressures of commercialization of some aspects of service delivery that is exacerbated the points above
- High costs of particular aspects of service delivery, in particular transportation costs that need to be taken into account when designing sustainable services
- Need to engage with governments and policy makers to help resolve these issues in partnership with all relevant stakeholders over the long term
Most of these issues are not unique to the African region; they have much in common with other parts of the world and echo the challenges faced by many HVP members. The broader challenge is to ensure that all people, no matter where they live, should be able to benefit from the recent developments in human genetics and genomics and, in particular, from supporting the growth in the number of quality jobs and services that are linked to knowledge-based economies. There is an increasing African diaspora linked to many leading institutions, hospitals and laboratories that can also be brought into a world-wide collaborative network to assist this development. There is a need for the Human Variome Project to assist by nominating a priority project to focus on the practical steps that can be taken to:
- oversee growth in the quality and quantity of curated databases by ensuring that best practices are followed in more countries, included those in the African region by consultatively developing all the necessary standards and procedures so that all relevant data can be shared in a harmonized manner to promote health service delivery;
- ensure that the harmonized storage, curation and sharing of the relevant variation information is financially and ethically sustainable;
- promote knowledge sharing among a formal network of professionals, including researchers, clinicians, bioinformaticians, counsellors, patients groups and health bureaucrats that will promote health services involving genetic and genomic components;
- ensure that all relevant ethical and regulatory frameworks and policies are developed and implemented to protect patients hand in hand with the development of the biotechnical systems and procedures; and
- give greater voice to those professionals from low and middle income countries working in human genetics and genomics in the development of standards, policies and solutions.
Tackling haemoglobinopathies like thalassaemias and sickle-cell disease, may prove to be an ideal entry point for many countries, to join others like Malaysia and China for example, that wish to develop infrastructure and expertise in other areas of human genomics and health service delivery.
Further opportunities for developing these collaborative efforts was raised through the presentation on Attracting EC Funding by the European Commission on their future funding priorities as set out in the forthcoming Horizon 2020 grants.
Challenges of finding data and interpreting data for clinical outcomes
Identifying data sources and enabling access in an efficient and coordinated manner is an on-going issue. There are various aspects to the problem and solutions require active participation from the full range of those directly involved in generating data as well as using it. Many example were shared of how to upgrade search interfaces, enhance linkages between linkage to multiple datasets using appropriately de-identified case matching, and how to integrate a genome wide database using LOVD3. The major questions of utility, quality control and sustainability of genetic variation databases is of concern to many in different parts of the world. Sharing knowledge on what works needs to be followed closely by more proactive efforts at networking and harmonizing activities. Working hand in hand with the local national or regional Human Genetics Society through a mechanism like an HVP Country Node can assist.
Recently, three NIH-funded efforts were aligned with the National Center for Biotechnology Information's (NCBI) ClinVar database under the collaborative Clinical Genome Resource Program (ClinGen). ClinGen is developing interconnected resources for the community to improve our understanding of genomic variation and optimize its use in genomic medicine. A unique aspect of ClinGen is that it represents a strong public-academic-private partnership that relies on the collaboration between NIH, academic and commercial genetic testing laboratories. The project includes the development of standards for variant interpretation as well as data submission and sharing. ClinVar, launched in April 2013, is a cornerstone of the project as it serves as the primary site for deposition and retrieval of variant data and annotations. As of February 1st, 2014 ClinVar contains 73,487 submissions across 18,702 genes (66,956 unique variants) with interpretations from OMIM, GeneReviews, 60 laboratories, and 23 locus-specific databases (LSDBs). The dataset includes 5454 variant submissions (2095 unique variants) from the Sharing Clinical Reports Project (SCRP) on BRCA1/2 and 4100 copy number variants from the International Standards for Cytogenomic Arrays (ISCA) Consortium. New policies and data structures are being considered to support controlled access to patient-level data. ClinGen is currently working with many laboratories and LSDBs to support robust mechanisms to share their data in an ongoing manner and increase the content of structured data and supporting evidence. Other parts of the project include computational and machine-learning approaches for identifying clinically relevant variants, and the development of expert working groups across many clinical domains to support consensus-driven evidence-based curation of genes-disease associations and genomic variant interpretations. Groups have already been formed in the areas of cardiovascular disease, hereditary cancer, metabolic disease, rasopathies, congenital muscular dystrophy, and developmental delay. The project is also interfacing with a large and diverse community of stakeholders including professional organizations, patient advocacy groups, regulatory agencies, research consortia and other projects from both national and international sites which is facilitated by working with the existing International Collaboration for Clinical Genomics (ICCG).
Thousands of new variants are being identified as a result of rapid advances in sequencing technologies. However, much of the data are stored in separate and sometimes private databases and so may be difficult to use in the evaluation of the clinical significance of variants. This is especially the case with rare variants. To improve access to this type of data, ClinVar (LINK SLIDES) maintains a freely available, public archive of human variation and its relationship to disease. The data can be used through a several means: interactively on the web; a monthly full release in XML format; and weekly summary files of genes and variants are also available for incorporation into analysis pipelines. Submissions include variants identified by direct testing in clinical or research labs, as well as reviewed variant-phenotype relationships from expert groups, such as InSiGHT and CFTR2, and professional societies, such as the American College of Medical Genetics and Genomics. In addition to the variant and phenotype, individual submissions may also provide a clinical assertion and evidence for that interpretation. The data model is flexible for many data elements, such that a variant may be defined by sequence or cytogenetic nomenclature; the phenotype may be a diagnostic term or features of a disease; and evidence for the interpretation may be structured as counts or provided as free text. For submitters who maintain their own website for variants, such as LSDBs, ClinVar links to the submitter's site for each submitted variant, allowing users who start at ClinVar an awareness of the LSDB's curated variants and access to more information on the variant that may be available at the LSDB. Each individual submission is accessioned and versioned, in the format SCV000000000.1, to allow the submitter to update their record as the interpretation of the variant is re-evaluated over time. ClinVar uses standard terminologies, such as those for variant nomenclature, phenotypes, and pathogenicity, to avoid data ambiguity and to promote comparison of information from multiple sources. ClinVar also adds related variant data, such as allele frequencies and HGVS expressions mapped across molecule types. While ClinVar staff members provide some curation of variants and phenotypes represented in ClinVar, clinical significance values are provided by submitters. As part of the submission process, ClinVar provides feedback to submitters. This feedback includes invalid HGVS expressions and submissions that conflict in clinical significance with an existing record for the same variant and phenotype which may warrant further curation. Submissions for the same variant-phenotype pair from different submitters are aggregated into a record that is accessioned and versioned in the format RCV000000000.1. Aggregation allows ClinVar to indicate when multiple submitters agree or conflict in the clinical interpretation of the variant, which can help clinical labs and curation groups to identify high-confidence interpretations as well as those that should be prioritized for curation efforts.
The International Society for Gastrointestinal Hereditary Tumors (InSIGHT) has established a committee (Variant Interpretation Committee; VIC) for the interpretation of sequence variants in the mismatch repair (MMR) genes associated with Lynch syndrome (LS). One of the major steps involved in this process has been the establishment of qualitative specific classification rules for the MMR genes, with the aim to improve the clinical utility of MMR gene testing. The 5-class variant classification system proposed by the International Agency for Research on Cancer (IARC) (LINKto website) was used to this purpose, since it links all classes to specific clinical recommendations. Multiple lines of evidence were required for class assignment and in order to classify a variant as pathogenic or likely pathogenic, (Classes 5 and 4, respectively), or as not pathogenic or likely not pathogenic (Classes 1 and 2, respectively), concordant evidence derived from both clinical and functional datasets had to be available. Variants with discordant information or with lack of either clinical or functional information were considered of uncertain significance (Class 3). The following specific points of evidence were considered:
- type of sequence variation;
- functional protein assays;
- mRNA assays;
- phenotype associated in compound heterozygotes for the variant under scrutiny and a clearly pathogenic variant in the same gene;
- presence of the variant on different haplotypes across LS families;
- co-segregation data and clinical phenotype;
- tumour molecular characteristics;
- population frequency;
- risk estimated from case-control studies.
Since interpretation of functional assays proved to be difficult and variable across committee members, specific supporting information and flowcharts have been developed to assist them. In addition, whenever available, quantitative multi-factorial analysis was used and the outcome compared to that of qualitative assessment. The classification scheme was modified by consensus to accommodate new data and inconsistencies over multiple classification teleconferences and face-to-face meetings. Overall, the rules were successfully applied to classify 2,360 variants lodged onto the InSiGHT database. These criteria provide a baseline for standardized clinical classification of MMR gene sequence variation that may be linked to patient and family management in the genetic counselling arena according to published guidelines. The VIC held a meeting in conjunction with HVP5, with an open invitation of meeting delegates to observe and understand the process used.
Locus-specific databases (LSDBs) that accept submissions of variants and their interpretation from diagnostic and research laboratories are subject to problems of duplicate entries, limited annotation of relevant clinical and experimental data, and variable quality of in silico analyses. In addition, ad hoc or generic criteria for interpretation lack refinement for gene specific interpretation, and do not draw on the specific expertise of clinicians and scientists directly working in the field relevant to the genes in question. As a result, interpretation of the same variant submitted from different sources can be discordant, leaving the database and the clinical field open to uncertainty, detracting from its utility for clinical application. The International Society for Gastrointestinal Hereditary Tumours (InSiGHT) has fostered the integration of all major mismatch repair databases into a single LSDB. InSiGHT has developed a coordinated approach to the authoritative interpretation of mismatch repair gene variants, engaging a range of gene specific experts, and drawing on comprehensive clinical and experimental information sourced from the published literature and from its own members.
The Committee applied a standardized classification scheme to constitutional variants in the Lynch Syndrome genes MLH1, MSH2, MSH6 and PMS2, reviewing all clinical and functional data available for the variants. All 2,360 unique sequence alterations were considered and classified. Assessment using validated criteria altered classifications for 66% of 12,006 database entries. Clinically important misclassifications were identified for unique variants initially submitted as not pathogenic (54 unique variants reclassified as pathogenic and 25 reclassified as likely pathogenic) and unique variants submitted as pathogenic (28 unique variants reclassified as not pathogenic, 16 reclassified as likely not pathogenic and 218 reclassified as uncertain clinical significance). Importantly, the consensus results have been disseminated online through the InSiGHT database and ClinVar. Consistent clinical management based on transparent evaluation is now possible for the 1,370 variants that are not obviously protein truncating (e.g. missense, single amino acid indels) based on their nomenclature. This large-scale endeavour demonstrates the value of multidisciplinary collaboration for curation and classification of variants in public locus-specific databases. The InSiGHT Variant Interpretation Committee (VIC) comprised of 40 experts in the mismatch repair field was tasked with variant review and classification on a pro bono basis. InSiGHT developed governance and support through incorporation, to protect the committee from possible legal challenge related to clinical use of its conclusions.
The Human Variome Project Australian Node (HVPA) has established systems and processes for capturing, curating, interpreting and sharing genetic variant information for diagnostic, treatment and research purposes. To support the accurate interpretation of variant information, linkage to clinical treatment and outcomes data from relevant health service providers is required. Unfortunately, the required clinical data is not collected systematically within the Australian healthcare system and linkage of such data is beyond the scope and resources currently available to HVPA. HVPA has partnered with BioGrid Australia, an independent not-for-profit organisation that provides a framework and infrastructure for data linkage and sharing that addresses patient privacy, data security, ethical issues and intellectual property concerns. Importantly, BioGrid's vision and mission closely align with that of the Human Variome Project. Furthermore, BioGrid has existing relationships with, and linkages to clinical data from, numerous healthcare service providers, research institutes and universities around Australia. While BioGrid's governance and legal framework, coupled with its existing infrastructure, provides the requisite platform to achieving national linkage of clinical data with the HVPA collected variant data, this requires time and significant stakeholder engagement to achieve. However, within the existing network of HVPA and BioGrid collaborators, we are currently undertaking a project to demonstrate the real potential value of data linkages for clinical and research purposes. A number of clinicians treating colorectal cancer at The Royal Melbourne Hospital are existing collaborators of both HVPA and BioGrid, providing an excellent opportunity to examine the potential benefits of data linkage. The pilot project is seeking to link de-identified, patient record level data from a range of sources with the goal of determining the depth and breadth of clinical data available for patients with specified variants at The Royal Melbourne Hospital. The secondary goal of this project is to determine whether, on review by subject matter experts, this clinical information can be used to support the determination of the pathogenicity of identified variants. By using existing clinical datasets and BioGrid's data linkage platform to integrate de-identified, patient record level clinical and genetic data, HVPA can efficiently build a national capability to capture, curate and interpret genetic variant information. Further leveraging BioGrid's online data access application system that incorporates scientific review with ethics committee oversight, HVPA can also ensure that data is shared with authorised users for approved diagnostic, treatment and research purposes. Other HVP Country Nodes are currently considering this model approach used by HVPA.
In summarising the meeting, Prof Garry Cutting, co-chair of the HVPI ISAC and HVPIL Board Member, indicated that annotating the human genome has many challenges. There has been a rapid increase in the number of genes associated with Mendelian phenotypes and an increasing number of variants for each gene - these come together and creates what he terms a 'perfect storm'.
To address this situation there is a need for secure databases that are well managed, able to produce good quality and useful information in an efficient and sustainable way. This means that laboratories in countries need to be connected, and exchanging their data. This can only happen if they are able to harmonise their operations and protocols. This means utilising similar processes and procedures, follow guidelines that ensure the above. This in turn leads to a collaborative working in partnership with all the key stakeholders. HVP members have seen these models work in relation to Cystic fibrosis and some inheritable cancers. Now is the time to extend these models into new disease areas by identifying key priority areas, having countries work together to solve problems in a practical and progressively extend their collaborative networks to include newcomers. This will be the challenge for the immediate future.