KOMP Phenotyping Conference

October 29-30, 2009

Hyatt Regency Hotel, Bethesda, Maryland

EXECUTIVE SUMMARY: At a Knockout Mouse Project (KOMP) Phenotyping Conference in Bethesda on October 29-30, 2009, a group of 25 researchers representing diverse scientific backgrounds expressed their unanimous enthusiasm and support for a proposed effort to conduct high-throughput and comprehensive phenotyping of the resource (targeted deletions for every gene in the mouse genome) being produced by the members of the International Mouse Knockout Consortium (IKMC). The researchers resoundingly supported an efficient strategy to first convert as many targeted ES cells into mice at centralized facilities that can then also quickly and efficiently conduct highly informative preliminary analyses (e.g., LacZ expression, stage of embryonic lethality, fecundity) at that same facility. As it is generated, the information would then be released to the community, with the mice and their germplasm being archived, for future distribution to the research community. The meeting participants stressed that the U.S. phenotyping effort should be coordinated with ongoing, established efforts in Europe, Canada, and Asia. As in these established phenotyping efforts, the U.S effort should adopt standard phenotyping tests, analyses, and examinations primarily on the basis of their reliability, power, and likelihood to reveal disease-relatedness and human clinical relevance. For the maximum likelihood of novel discoveries, the effort should initially be focused on studying poorly understood genes. Results from a primary (baseline) phenotyping screen conducted at centralized sites on as many genes as possible should be used to inform secondary (challenge) and tertiary (specialized) phenotyping. Phenotyping data and information should be made readily available without restriction or moratorium through web portals that are user friendly and well advertised. These recommendations were highly correlated with results from a survey and Request for Information (RFI) requesting input from the US national research community (see page 5).

KOMP Phenotyping Conference Report

On October 29-30, 2009, twenty-five scientists representing multiple scientific disciplines from across the United States were invited to participate in an NIH supported workshop organized by Dr. Kent Lloyd, of the University of California, Davis. Also in attendance, and contributing substantively to the discussions, were leaders of the IKMC programs, the KOMP panel of scientific consultants, and NIH program staff. The purpose of this meeting was to discuss the development of a strategy to phenotype the IKMC resource. Attendees were informed and updated on the KOMP production effort in the context of complementary international efforts on mouse mutagenesis (EUCOMM, NorCOMM, and TIGM) coordinated through the IKMC. Not surprisingly, attendees learned that ongoing efforts to phenotype IKMC-generated mutant mouse lines by MRC-Harwell, The Sanger Institute, Helmholz Zentrum Munich, and the Asian Phenotyping Consortium have revealed previously unknown and unrecognized gene function and disease associations, clearly demonstrating the utility of this endeavor. In addition, the attendees were presented a summary of comments from more than 350 respondents to a recent, publicly released RFI and a directed survey (see attached) that indicated broad community enthusiasm for a comprehensive effort to phenotype the IKMC resource that would directly facilitate investigator-initiated research and translational science. Further, the attendees heard from scientists who had developed and applied phenotyping screens to ENU-induced mutants. With this extensive background information, the attendees unanimously concluded that a well coordinated and strategically executed phenotyping plan could directly inform research and discovery on human disease, developmental disorders, and hematological, immunological and behavioral abnormalities, among others. Upon completion of this update, the attendees were asked to provide the NIH with informed input and recommendations on goals, approaches, scope, and logistics of conducting a hypothesis-generating effort to phenotype the KOMP resource that has the greatest potential for leading to hypothesis-driven research to improve human health. Based on the model in which the current NIH KOMP is able to secure continued funding of at least $11M per year to phenotype 1000-1500 mouse lines over the next 5 years, the attendees made the following recommendations:

1) Gene selection: Gene knockouts to be converted from targeted ES cells into mice for high-throughput analysis should be prioritized from the IKMC gene list for discovery of new phenotypes, not triaged for confirmatory, or predictable, phenotyping. Selection of genes should be based on a number of factors, including disease-association (e.g., OMIM), specific expression patterns, and/or known or suspected gene function. Further, gene selection should also consider human clinical and disease relevance so as to provide compelling justification for continued support of funding these initiatives across multiples IC’s. Several gene families from which genes could be selected for priority phenotyping include, but are not limited to, nuclear envelope-associated proteins, signaling proteins, transcription factors, DNA repair genes, nuclear-encoded mitochondrial proteins, and extra-cellular matrix proteins. Additionally and importantly, consideration should be given to the selection of a large proportion of genes that are understudied, poorly characterized or annotated, and for which there are few if any publications. Phenotyping this category of genetically targeted mice has the greatest potential for new discoveries. Data resulting from these studies are likely to provide preliminary data for competitive, investigator-initiated research. To prevent unnecessary overlap and inefficiencies, gene selection should be coordinated with ongoing international phenotyping efforts

2) Mouse production and preliminary analysis: It is evident that the bottleneck limiting the breadth and depth of any proposed KOMP phenotyping plan will be generating live mice from ES cells. Therefore, the discussants strongly recommend that mice be generated as soon as possible for as many genes as possible. The importance of quality control, and the need for sufficient capacity, operational efficiency, and specialized expertise, strongly supports the recommendation for centralization of the mouse production effort. Centralization will also enable critical “upstream” analytical data collection that will contribute to more downstream logistical choices for phenotyping. In this context, the discussants recommend the mouse central production facility perform the following phenotyping in house: LacZ expression analysis of the mutant allele in heterozygous embryos (n=3, E12.5 days post coitus [dpc]) and adult animals (1 male & 1 female, 8 wks); discern developmental competence vs. embryonic lethality by genotyping litters from heterozygous parents and assessing embryo development in time-pregnant females at E12.5 dpc; and determine fecundity. Additionally, live mouse lines should be made readily available for distribution to the research community, and converted and archived as cryopreserved germplasm for preservation and protection, thereby ensuring reliable and continuous access to the resource.

3) Analysis: Because of the value placed on testing protocols that reveal undiscovered phenotypes, the phenotyping approach should employ a primary screen of standardized baseline (naïve) tests, examinations, and analyses applied equally to cohorts of mice consisting of 7 male and 7 female homozygous or, if the resulting homozygous knockout mice are not viable, heterozygous mice. The tests in this primary screen should be chosen on the basis of numerous characteristics, including reliability, sensitivity, and specificity, diagnostic power, clinical relevance, historical hit rate, and an optimized cost:benefit ratio. Further, development of the primary screening strategy should be based on and influenced by current phenotyping activities (e.g., MGP pipeline at The Sanger Institute, EmpressSlim at MRC Harwell, and the phenotyping pipeline at Helmholz Zentrum Munich). Finally, centralization will reduce expenses and inefficiencies associated with shipping live mice, and ensure the reliability and minimize the variability of testing outcomes. In contrast, blood, tissues, and other specimens can be shipped off-site for analysis more easily and with less influence on results than shipping live mice. Shipping biological samples will increase access to specialized phenotyping and increase community involvement in the project.

To these ends, to the extent possible, and in approximately rank order of priority, the following tests, analyses, and examinations should be considered in a primary phenotyping screen applied to all genes:

• Body weight (weekly)
• Visual observations/modified SHIRPA
• ABR/auditory (acoustic startle)/olfaction
• Open Field/Automated activity monitoring
• Grip strength
• Ophthalmologic/retinal exam
• Body morphology/laser scan
• Echocardiography
• Acoustic startle
• Hot plate
• Radiology/DEXA/MR
• ipGTT/insulin
• Hematology/Urinalysis/Clinical Chemistry (amino acids, liver enzymes, leptin)
• FACS (immune panel)
• Platelet aggregation/hemostasis (blood coagulation)
• Basic necropsy/organ weights/tissue blocks
• Additional tests that should be considered are expression profiles (e.g., microarray), mass spectrophotometry, micronuclei, and plethysmography.

The following challenge tests should be considered in a secondary phenotyping screen applied to all or a subset of genes that have undergone primary phenotypic screening. It should be noted that LacZ analysis should be performed to determine expression of the mutant allele in response to challenge:

• Diet (fat and CHO)
• Embryo challenge
• Immune (pathogen, antigen, allergen, cytometric monitoring)
• Cancer
• Aging / genetic sensitization

Thereafter, further analyses that reveal gene pathways, disease mechanisms, gene family analysis, treatment modalities and other follow up studies should be considered for specialized phenotyping in individual research laboratories.

4) Data access and presentation: As the many international mutagenesis efforts are quickly advancing in the production of targeted alleles in ES cells, the breadth and depth of data and information being generated is fast becoming enormous. To properly manage this information, a number of websites and web portals are being developed, including IKMC, IMPC, and individual center websites. The discussants strongly recommended that the creation of a uniform, user-friendly web portal that integrates all of this data and information, making it freely accessible without holdback or moratorium soon after it is generated, is critical to optimize utilization of IKMC products for research. To the extent possible, phenotypic data and information should be presented in standard formats, be highly searchable, and be integrated broadly with genome browsers, public databases (e.g., MGI, Wikipedia) and other informatics tools (e.g., Biomart, etc.). Discussants promoted the value and importance of customer service with respect to finding and accessing phenotyping information, including automated updates on production and phenotyping pipeline statuses, availability of ES cells, mice, and test results.

In summary, any effort to phenotype KOMP resources should be fully complimentary to, and closely coordinated with, current and ongoing phenotyping efforts by other members of the IKMC. To that end, active engagement with the international mouse phenotyping consortium (IMPC) is recommended. The funding mechanisms used to support this effort should be sufficiently flexible to accommodate the varied approaches to phenotyping KOMP and other IKMC mutant mouse lines. Finally, the NIH should establish mechanisms to require researchers to deposit phenotyping data, or utilize phenotype data to generate additional research findings, into publicly accessible databases within a reasonable time frame for the benefit of the entire research community.

SUMMARY OF SURVEY INSTRUMENTS SEEKING INPUT ON KOMP PHENOTYPING

1) KOMP Phenotyping Survey

Key Dates: Released: September 8, 2009, Closed: October 15, 2009 (extended from October 1)

Survey recipients: Survey was emailed to 2517 individual persons who established user accounts at www.komp.org. A reminder was sent at the end of September, and the survey close date was extended 2 weeks. Survey recipient were reminded that their responses were submitted anonymously.

Description (as written in the survey instrument): For this survey, assume that a high throughput phenotyping effort will go forward on all KOMP genes over a ten year period. The goal of the survey is to ascertain what information “area-specific” researchers would need from a primary phenotype screen to enhance research efforts and advance the field of study in their area of interest. Thank you for participating in this survey...it will ensure your voice is heard!

Results:

-High response rate (13%)…338 responses.

-Many good suggestions received…most were very enthusiastic about an effort to phenotype the KOMP resource.

-¾ of the responders said that phenotyping would motivate them to use KOMP mice.

-More than half suggested tests to assess behavior, metabolism, obesity, immune system, embryo development and early embryonic death, cardiovascular, and peripheral vascular system.

-They want more tests with immune challenge and for cancer and embryogenesis.

-Three essential tests are in immunology, metabolism, and behavior.

-¾ of responders prefer challenge tests over naïve or challenge tests alone.

-Top 5 baseline tests are behavior, exercise, development/expression, growth/morphology, and general observations.

-Top 5 challenges are diet, pathogen, antigen/allergen, carcinogen/cancer, and drugs.

-A little more than half of responders prefer triage approach to phenotyping compared to subjecting all mice to same tests.

-¾ of the responders are willing to annotate phenotyping results.

2) Request for Information (RFI): Phenotyping Knockout Mice (NOT-HG-10-005)

Key Dates: Released: September 24, 2009, Closed: October 15, 2009

Issued by: National Human Genome Research Institute (NHGRI); National Center for Research Resources (NCRR); Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD)

Description (as written in the survey instrument): The purpose of this RFI is to seek input and feedback from the scientific community in order to inform and guide future large-scale phenotyping of knockout mice. Specifically, the NIH is seeking advice and recommendations on the scope, depth, and breadth of a phenotyping effort, and to provide rationale and justification on how this effort would benefit research.

Information Requested (as written in the survey instrument): For the purposes of this survey, assume that a high throughput phenotyping effort will go forward on all genes over a ten year period. The goal of the survey is to ascertain what area-specific information researchers would need from a primary phenotype screen to enhance research efforts and advance the field of study in their area of interest.

Results:

-Low response rate…24 responses.

-Many good suggestions were received…most were very enthusiastic.

-All responders agreed that phenotyping effort was valuable and beneficial.

-~65% wanted more baseline than challenge tests.

-Most were interested in phenotyping for behavior, development, neuro, seizure, heart, and fertility, as long as the results were clearly evident.

-~¾ valued challenge tests.

-Top 5 baseline tests were pathology, baseline, behavior, fertility, growth/morphology, and general observations.

-Top 3 challenges tests were diet, pathogen, and carcinogens.

-~¾ preferred conducting all assays on all mice, rather than a triage approach.

-Half of respondents estimated costs for phenotyping a mutant at $10K-$100K.