Case Study

National Institutes of Health, National Center for Research Resources: The Initiative to Link Animal Models to Human DIsease (LAMHDI) – A Software Treasure Map

Key Findings

  • Using open-source software ensures the government and the researchers who use LAMHDI will not be held hostage to a proprietary system
  • An effective outreach strategy is key to fostering widespread acceptance and use
  • Partnering with scientists from the beginning ensures our work meets the needs of both information users and providers
  • Building on the work of our partners and others, both in data collection and curation and in software functionality, helps drive success

Challenge

While animal models are central to the effective study and discovery of treatments for human diseases, finding appropriate models for every stage of research is still a challenge. Scientists are more interested than ever in the three Rs of animal model research: refining, reducing, and replacing models where possible. But cross-species information is hard to find and hard to evaluate, because each model database—where databases exist at all—is different. And the literature often does not give the kind of detail researchers need: What are the alternatives? Where can the model be accessed? LAMHDI is intended to gather and organize information about extant animal models across species and strains, and make it available and searchable through a Web interface.

To accomplish this, the initiative for Linking Animal Models to Human Disease (LAMHDI) was initially funded to create a resource for integration and sharing of data and information about animal models.

The initiative consists of two components:

  • The development of a directory of available human-disease models that will assist researchers in finding information about those models, and support experts who are providing and organizing such information.
  • The creation of a knowledge-based portal that provides a single point of contact from which all models may be accessed and processed.

The NIH hopes that the lessons learned in the LAMHDI project, and the processes and structures created for collecting, organizing, improving, and sharing models and data, help set a standard for successful knowledge environments.

Solution

LAMHDI is a Web-based resource to allow users to integrate data and information about animal models to enable the animal-model and human-disease research communities to identify appropriate resources and apply them in their research. Typically, researchers seeking appropriate animal models for their investigations into human diseases settle on models that are familiar, because exploring options across species, kinds, and laboratories is tedious and requires specialized knowledge of nomenclature as well as the ability to compare dissimilar data. LAMHDI is making that quest easier by automatically finding the natural links from one model, one species, or one source to all the others. Initially LAMHDI is focusing on two animal species, mice and zebrafish. LAMHDI will then expand to other species and along other dimensions such as microbes or tissues.

LAMHDI does not create the data it offers to users; that is collected and curated by scientists worldwide. LAMHDI’s role is to translate from one data system or data structure to another, so scientists can search across existing databases and find relationships that will help inform research.

TCG is skilled in building centralized resources for technologists, informatics practitioners, and healthcare professionals and brought our expertise to this project. Specifically, we:

  • Combined and melded best practices from the PMBOK, CMMI and RUP to ensure project management efficiency, timeliness, and quality
  • Partnered with scientists from the beginning, to ensure that our work would meet the needs of the users and providers of information
  • Worked with scientists and researchers to develop use cases and UI requirements for the LAMHDI web site
  • Involved scientific users in planning for ongoing work to build on their knowledge while engaging their interest and involvement
  • Built on our partners’ work to take advantage of their efforts, both in data collection and curation, and in software functionality
  • Mapped data sets from multiple animal model databases to provide a single cohesive search process and engine
  • Worked with NIF and NeuroLex ontologists to develop synonym mechanisms for improving the completeness and accuracy of model searches, and
  • Built LAMHDI using open-source software so that the government and the researchers who use LAMHDI would not be hostage to a proprietary system.

We continue to garner feedback from the scientific community to improve LAMHDI and ensure relevance and timeliness of the model data

We did not limit our activities to the technology and data. We recognized that a horn not tooted is a horn not heard, and we designed a communications strategy that would build on our partners’ knowledge, reputations, and leadership in the human-disease research communities. As a result, LAMHDI usage is growing through promotional appearances at scientific meetings and through publications and presentations by our scientific partners, and also through guerilla marketing, as scientists mention its utility to one another in casual conversations—conversations initiated by the scientists who have been involved as advisors and contributors.

Results

  • Prototype site was delivered in 12 months, and exactly on schedule.
  • The prototype site met NCRR’s aims but, with NCRR’s agreement, we re-designed it in preparation for the production roll-out. In four months we provided greater search capabilities and better integration of existing and new databases and data. This shows our ability to identify opportunities for improvement in a pilot system and quickly implement significant improvements for production
  • We involved additional scientists and new technologies, going beyond our original plans to better serve the broader community
  • NCRR happiness with our work resulted in one third again as much money as originally contracted

Taxpayer Savings: $10.75 million

For every project, TCG tracks the savings generated by innovation in methods and technologies. LAMHDI has saved the US taxpayer over $10.75 million. Here are some ways that we generated these savings:

  • We save the cost of literature searches, which we estimate to cost about $250 each—double the cost of a search by a professional searcher. A 2010 paper identified 1500 experiments using animal models for ischemic stroke research (see here). Assuming that ischemic stroke research accounts for 2% of the human-disease research using animal models (admittedly a WAG), that brings the number of experimental designs using animal models to 75,000. If each of those research projects saves one-tenth the cost of the literature search to find the right model (literature searches still have to be performed), then LAMHDI will have saved $1,875,000 a year, or $3,750,000
  • Universities are not expert in project management; university projects are geared toward research, where the outcome is unknown. Projects, on the other hand, are designed to deliver a specific product or service. By melding TCG’s project management expertise with input from researchers at universities, American taxpayers saved some $1-million (see here) because of the average 88% return on investment in professionally managed projects.
  • By reusing code developed for TCG’s NITRC project, and curation work done by our research partners, TCG saved the cost of those projects: $500,000 in curation activities; $3.3-million to build the zebrafish database; and $850,000 a year for the mouse database. So LAMHDI saved $1.3-million a year for its first two years, plus the $3.3 million that it would have taken to recreate the zebrafish database, a total of almost $6 million in its first two years.