Home > Strategic Planning > National Eye Institute Workshop to Identify Gaps, Needs, and Opportunities in Ophthalmic Genetics [NEI Strategic Planning]

National Eye Institute Workshop to Identify Gaps, Needs, and Opportunities in Ophthalmic Genetics [NEI Strategic Planning]

National Eye Institute Workshop to Identify Gaps, Needs, and Opportunities in Ophthalmic Genetics

Meeting Summary and Recommendations

Airlie House, Virginia
June 4-5, 2009

The meeting consisted of three presentation sessions: 1) Ophthalmic Disease Conditions, 2) Biological Systems, and 3) Approaches and Methodologies.

Drs. Janey Wiggs and Elaine Mardis moderated a final session in which a consensus of Crosscutting Needs was developed as follows:

Databases: Although the NIH maintains the database of Genotypes and Phenotypes (dbGaP), more dynamic and comprehensive electronic databases could be developed for all NEI-supported samples. Current platforms are inadequate to handle, store and backup the large volume of new data. First a comprehensive evaluation of what is currently available and what the vision community needs would be useful. A workshop focusing on databases could then assess specific needs and opportunities.

Tissue/cell line/DNA Repository: In addition to storing samples, a central repository would organize DNA testing results from multiple laboratories. A repository could also store cell lines created from human subjects to leverage the investment in high quality phenotyping. The NEI could develop a resource to assist researchers in getting access to, storing, and handling post-mortem ocular tissues.

Phenotype/Genotype Centers: To facilitate large-scale genetics research, 10-15 centers throughout the US could be formed to conduct phenotyping, reduce marginal costs of genotyping, create animal models based on human genetics, establish databases, and coordinate bioinformatics. Major phenotyping centers could provide uniformity as protocols are harmonized across centers. DNA samples will be screened for relevant genetic information and deposited in the database (above). Relevant tissues could be banked for future use.

Collection of Super controls for ocular studies: Phenotyping centers could develop a reference sample of comprehensively phenotyped controls. Cell lines/DNA samples will be collected and maintained (e.g., for at least 1000 controls) from each of the phenotype/genotype centers. Control DNA and phenotype information would be available to investigators. It is critical to harmonize what phenotypic data are collected, how to collect it, and how to measure it.

Assembling large cohorts: Super controls and large centers require large patient populations. Collaborative research networks of clinicians across the US such as the Diabetic Retinopathy Clinical Research Network are a well-tested model.

Pharmacogenomic adverse events: Support next generation sequencing for pharmacogenomic ocular adverse events (such as plaquenil toxicity, topomax angle closure glaucoma).

Training: High-throughput technologies yield huge data sets. Collecting, storing, processing, and analyzing data requires a new generation of biomedical researchers with computational and statistical expertise. Training the next generation of vision researchers in statistical genetics and bioinformatics is critical.

Each of the following three sessions included a presentation followed by a discussion during which one or more specific needs and opportunities were identified (shown in italics). NOTE: These reflect the opinions the presenter although often, other participants contributed.

Session 1: Ophthalmic Disease Conditions

(Donald Zack, Moderator)

Strabismus and Amblyopia
(Elizabeth Engle)

Strabismus is the only human genetic model for axon guidance disorders. Early detection is critical for strabismus and amblyopia since treatments are more effective when started during childhood. Genome wide Association Studies (GWAS) have not been done. Strabismus can be divided into well-defined clinical subtypes. One strategy to ensure studies are adequately powered is to conduct large studies, followed up with subtype analysis. Another way to maximize resources could be to combine GWAS on strabismus and refractive error.

  • Conduct case-controlled GWAS for strabismus/amblyopia in multiple centers, enrolling 250 probands/yr to identify genes associated with common strabismus.

Retinal degenerations
(Michael Gorin)

Family-based linkage studies and targeted gene re-sequencing will reveal causative mutations, and additional molecular genetic work will need to be done to interpret the downstream functional consequences.

  • Identify modifier genes for retinal degeneration.
  • Enhance molecular diagnostics, including mutation and causative gene discovery (e.g. exomic sequencing, family-based studies)

Age-related Macular Degeneration
(Anand Swaroop)

Genetics may allow patient stratification based on expected treatment response (i.e. risk-assessment screens/personalized medicine).

  • Conduct genetic analysis of retinal aging.
  • Conducting meta-analysis of GWAS across ethnic groups may provide access to molecular pathways of AMD.

GWAS have generally not included African Americans because they have low rates of AMD. Studies of African American AMD populations could reveal missing protective factors for AMD.

(Louis Pasquale/ Janey Wiggs)

Clinically, glaucoma is a heterogeneous disorder with multiple interacting pathways. Studies need to concentrate on clinical endpoint phenotypes and endophenotypes to facilitate genetic analysis.

  • Conduct GWAS in POAG for moderate effect genes and gene-environment interactions.
  • Conduct GWAS for psuedoexfoliation glaucoma.
  • Explore a joint study with Japan to study normal tension glaucoma.
  • Apply next generation sequencing to confirmed glaucoma loci.
  • Use mendelian/early-onset glaucoma to shed light on mechanisms of common glaucoma.

Long-term goals include identification of major genes and modifiable risk factors causing incidence of glaucoma.

(Terri Young)

Family based studies have revealed 20 loci for myopia. Longitudinal studies such as the Multi-Ethnic Study of Atherosclerosis (MESA) have genetic and some refractive data, but in adults, refractive error information is confounded by lens opacity, so data must be collected with biometry.

  • Perform GWAS for longitudinal clinical trials like COMET.
  • Conduct meta-analyses of refractive error on GWAS results, including international studies.
  • Develop biomarker profile for severe morbidity.

It is important to standardize biometric measurements for refractive error. Studies should collect environmental data, including factors such as education level.

Lens and Cornea
(Janey Wiggs)

Genetic disorders can cause opacification of the lens and the cornea. Genetic-based interventions must improve upon existing surgical interventions.

  • Use genetics to risk-stratify patients more susceptible to surgical complications.
  • Conduct genetics studies of Fuchs endothelial dystrophy and keratoconus to identify modest effect genes/modifiable risk factors.

Session 2: Biological Systems

(Jerome Rotter, Moderator)

Aquaporins (Alan Verkman) and Connexins (Daniel Goodenough)Mutations in Aquaporins cause many eye conditions (lens/corneal cataract, elevated IOP, tear secretion). Different Connexin mutations cause different cataract phenotypes in humans.

  • Create repository for AQP transgenic mice.
  • Encourage small-molecule discovery of AQP modulators.

Mouse Models
(Simon John)

Good animal models for complex diseases, especially aging-related models, are needed. A useful new resource being developed is the community cross where homogeneous lines are crossed to produce reproducible results in heterogeneous backgrounds. Other systems such as zebrafish, dogs, and tree shrews will be useful for studying certain conditions and testing potential therapies.

  • Humanize mouse models (replacing mouse gene variants with their human homologs).
  • Establish centralized basic mouse resources (such as sperm banks, ES lines, lines of mice with cre alleles, well-characterized lines with cell types marked by fluorescent proteins, point mutation resources).
  • Establish animal phenotyping centers to identify and validate disease models.

Stem Cells
(Robert Lavker)

Adult stem cells (ocular limbal cells or oral mucosal cells) have been used for transplantation in ocular epithelial cell deficiencies. Limbal stem cells can be deficient due to congenital, autoimmune, or acquired causes. Retinal regeneration requires a population of committed precursor cells; research identifying cell surface biomarkers would facilitate precursor isolation.

  • Identify markers for stem cells.
  • Define the critical components of the stem cell niche to improve their regenerative capacity and expedite expansion.
  • Standardize methodologies for ex vivo transplantation.
  • Establish communal GMP (good manufacturing practices) facilities for stem cells that meet FDA requirements so stem cell-based therapy products can be used in humans.

Session 3: Approaches and Methodologies

(Margaret Pericak-Vance, Moderator)

Statistical genetics and study design
(Alexander Wilson)

Different study designs and methods are required depending on the size and frequency of the effect in study (a priori, the size of the effect/mode of inheritance are unknown). Effect size is more important than frequency. Use linkage analysis and intra-familial tests of association in families to identify rare alleles with large effects, and population based studies to identify common variants.

  • Foster large scale collaborations that use a high-density SNP genotyping panel. Large-scale studies use harmonized phenotype definitions, genotype platform, and analysis methodology.
  • Genotype existing samples only when there are high quality phenotype data
  • Begin incorporating sequence variants in future studies (including GWAS).

Population studies–Ocular Phenotypes
(Rohit Varma)

  • Harmonize disease definitions (anatomical, physiological, environmental), but standardize phenotype/biomarker definitions, preferably with objective measurements (while keeping primary data in case definitions change later).
  • Design trials that target persons at greater genetic risk.
  • Collect genomic data in current and planned clinical trials.

Novel phenotypes will depend on developing advanced diagnostic systems such as spectral-domain Optical Coherence Tomography (OCT), non-invasive fundus oximetry, and 24-hr intraocular pressure (IOP) sensors.

Genome Wide Association Studies (GWAS)
(Lindsay Farrer)

GWAS are commonly case-control studies but may include case-only and family-based approaches. GWAS may be modified to address specific disease hypotheses, such as using a systems or pathway approach. In a genetically isolated population with less heterogeneity, genes can be found with lower sample sizes; genetically diverse populations yield more robust and generalizable results. Studies can be simplified by looking at heritable endophenotypes.

Next Generation Sequencing
(Elaine Mardis)

In addition to improvements in speed and data outputs, new sequencing methods are based on single molecules in flow cells; therefore they can be used for DNA quantification, detection of rare variants, and have enhanced resolution and dynamic range relative to microarrays, without certain inherent biases. However, disadvantages include the shorter read lengths (35-500 base pairs), which introduces systematic and accuracy errors, as well as bioinformatics challenges. Applications include comparing disease vs. control patterns in DNA methylation, protein-DNA binding, and non-coding RNA discovery.

Costs: currently, it costs $120,000 to sequence a genome, but predictions for the end of the year are about $50,000. Sequencing just the exome costs $12,000. Costs come down as read length (and data output per run) goes up and also depend on the market (competition between ABI and Illumina).

  • Long range goal: Develop platforms to use Next Generation Sequencing as a diagnostic tool for mendelian genes.

Proteomics (Reid Townsend)

Proteomics is much slower than genomics but new technology using peptide instead of protein-level analyses is improving speed.

Bioinformatics (Terry Gaasterland)

Bioinformatics can be used to model 3D structures of proteins to find functional sites, analyze gene expression patterns, construct regulatory models, compare genomic sequences, and help to provide network/pathway interpretation (for example in ocular oncology).

  • Tap into resources and tools by piggybacking on the NIH ENCODE project (ENCyclopedia Of DNA Elements).

Back »