The US stimulus package has created numerous opportunities for bioinformaticists — through dedicated award opportunities as well as supporting capabilities on non-computational programs.
When Congress passed the American Recovery and Reinvestment Act in March, the National Institutes of Health gained an additional $8.2 billion in funding to disburse over the next two years. Among a number of other funding mechanisms, NIH created two new programs — the NIH Challenge Grants in Health and Science Research and the Grand Opportunities awards — targeted at awarding these short-term funds. Each program has a budget of $200 million.
NIH's share of the Recovery Act funding "is a real boon for bioinformatics," Karin Remington, director of the Center for Bioinformatics and Computational Biology at the National Institute of General Medical Sciences, told BioInform via e-mail.
"Computational tools, data integration and analysis infrastructure, and data visualization have become critical components of the biomedical research endeavor, yet are quite expensive and time consuming to develop," she said. "The Recovery Act funding provides a great opportunity for the bioinformatics community to approach their work much more aggressively, whether scaling up their existing efforts with additional personnel, or beginning new work that previously seemed beyond the reach of our traditional funding mechanisms."
Two NIH institutes that tend to provide the largest share of bioinformatics funding, NIGMS and the National Human Genome Research Institute, have identified a number of funding opportunities targeted directly at informatics development. Information on NIGMS Challenge Grants is available here, and NIGMS Grand Opportunities are available here. NHGRI Challenge Grants are described here, while NHGRI Grand Opportunities are here.
"Many of [the Challenge Grants] have components related to informatics and computation," while some of the GO projects have "significant informatics pieces," NIGMS Director Jeremy Berg told BioInform. NIGMS expects to receive approximately $507 million in ARRA funds over the two-year period, but it has not disclosed how much of that might go to bioinformatics projects.
The Right Grant for the Right Project
"One of the big challenges [for applicants] is to find things you can do in two years that aren't just going to be ramping up for the interesting bits," Berg said.
As part of their application, scientists must outline what they plan to do with their research after two years, he said. "Either in two years you have accomplished what you've needed to do or there's a sustainability plan," Berg said. For example, a project to develop software will entail having robust software as the deliverable, he said.
Given the NIH-wide "huge interest in sequencing," Berg said a project might involve two years of data collection to offer a "wonderfully rich data set to a broad community," and researchers can then use normal NIH funding mechanisms to perform sequence analysis and data mining.
Indeed, officials at NHGRI and NIGMS told BioInform that scientists should not lose sight of the traditional grant opportunities and should allocate their projects accordingly. "The success rate will undoubtedly be higher, and you get four or five years worth of money instead of two," he said.
In addition, the NIH considers the areas being funded by the Challenge grants as "jumpstart funds" devised for research that will "quickly advance" science in "significant ways" by addressing, for example, "knowledge gaps" or creating new technologies.
[ pagebreak ]
Some scientists interviewed by BioInform said they wondered if the short deadline might lead to funding for less innovative projects or hurt chances for researchers to win funds for projects that require a longer development roadmap than two years.
Researchers seem to have questions regarding whether they should either continue working on previously started grant submissions that are geared toward a longer-term project, try to convert those projects into short-term programs, or to try to quickly devise a shovel-ready project for the two-year time-frame.
Several NIH program officers who requested anonymity said they are putting mechanisms in place to assure continued evaluation despite the time pressures.
Berg said that the Challenge Grant application form has been downloaded "well north of 10,000" times for 200 awards. "If we get an additional 15,000 [applicants] for 200 awards, then the success rate will be very low."
Berg expressed concern about a possible "huge mismatch" between the number of applications and the funds available — which he said represents another reason not to "forget about the normal [funding] mechanisms."
At NHGRI, Lisa Brooks, program director of the Genetic Variation Program, and Peter Good, program director of Genome Informatics, told BioInform via email they are not concerned about the quality of incoming projects, but echoed Berg's concern about possible low funding rates for the ARRA-related grants.
Brooks pointed out that scientists should remember "there are also opportunities to enhance current projects in valuable ways" through traditional funding mechanisms.
Meantime, Good pointed out that the grants may also be "an opportunity" for "different ideas that may not come up through the normal R01 channels."
For their part, Brooks and Good oversee a number of ARRA NHGRI Grand Opportunities that are relevant to bioinformaticists. One such opportunity covers programs designed to develop and apply new statistical and computational data-analysis methods for DNA-sequence, variation, GWAS genomic function, chemical biology, and other -omics data sets.
"We do not want to have the $1,000 genome that requires the $500,000 analysis," Brooks said. "The sequence platforms are changing rapidly, and other types of genomic data collection are also providing huge amounts of data that require analysis and integration."
Describing funding opportunities that cover statistics, she said there are "many" experimental-design issues that are statistical, such as the number of samples to examine in a genome-wide association study to detect variants of certain allele frequencies and allelic effects of a certain size, and how to take proper account of linkage disequilibrium when evaluating the significance of genome-wide associations, she said.
A separate NHGRI Grand Opportunity Brooks oversees is designed to develop a sequencing data software pipeline that will be passed to the NCBI, whose task it will be to implement the software and assess its data quality and call variants.
As Brooks explained, the 1000 Genomes Project has set up a pipeline for human resequencing data, taking in sequence reads and putting out variants with likelihoods. "This sort of pipeline should not be just project-specific, but should be run centrally from NCBI so [that] all such data can have the opportunity to be run through it and processed in a standard manner," she said.
'Bang for the Buck'
Some scientists developing web services or software told BioInform they are concerned that their projects might be left hanging after two years of funding, but Brooks said that any software that is made publicly available should last beyond a project's funding period.
The maintenance of any resource is always "a tricky problem" given the tension between funding resources and funding new science, Good said, adding that the NHGRI "will support resources that have a demonstrated impact on the community."
[ pagebreak ]
"What we always struggle with is how to determine which tools merit continued support, and also how to measure impact," he said. He also referred scientists to the ongoing NIH Biomedical Information Science and Technology Initiative, or BISTI, program, which is designed to maintain biomedical software.
Good oversees a GO project to develop a data-analysis and -coordination center for -omic data related to cancer, including those from the Tumor Sequencing Project and the Cancer Genome Atlas.
The GO topics that NHGRI put forward are "where we think the scientific community could get the biggest bang for the buck in a reasonable time frame," Good said.
"I know we'll be seeing some very innovative, exciting ideas," Laurie Tompkins, chief of the Genetic Mechanisms Branch in the Division of Genetics and Developmental Biology at NIGMS, told BioInform via e-mail.
Tompkins oversees a Challenge project on structural analysis of macromolecular complexes — grant 06-GM-101 — that is looking for technological innovations that are "either significant modifications of existing technologies or entirely new approaches" to facilitate analysis of macromolecular complexes.
She said the projects can include computational approaches, but that she has received "relatively few inquiries" from scientists in computational areas.
A typical NIGMS GO program relevant to bioinformatics is one that seeks to harden existing software and/or databases to create robust products for dissemination in the form of "professional-grade" Web-based resources.
In terms of NIGMS Challenge Grants, one in particular — 06-GM-103, overseen by Peter Preusch — seeks to develop predictive methods for molecular structure and ligand recognition. "The value of computational approaches to predicting macromolecule-ligand interaction is well known, as are the weaknesses of the current software for this purpose," Preusch told BioInform in an e-mail this week.
Any number of possible improvements can be envisioned, he said, adding that "we are interested in the investigators' ideas about what are the limiting factors and ways to overcome them."
Among other areas of interest are those designed to create tools to help predict "the absolute binding constant for any macromolecule-ligand interaction" and "correctly modeling macromolecule-macromolecule interactions," Preusch said.
In the area of data integration, one NIGMS Grand Opportunity is about integrating pharmacogenomics data with electronic health records. This project is intended to develop "a paradigm to use genotypes to predict drug responses in a real-world setting with the goal of safe and effective use of medications, program officer Rochelle Long told BioInform in an e-mail.
"There remains an art to sifting through the information computationally," she said. "We have to start somewhere and it's time to try to find the low-hanging fruit examples as proof-of-principle."
According to the project description, "large-scale pilots" are needed to demonstrate, for example, "the value of automated methods for computational definitions of health and diseases status."
NIGMS and NHGRI are certainly not the only NIH institutes with ARRA-related programs of interest to bioinformaticists. The National Cancer Institute lists several informatics programs on its Challenge Grants site, as does the National Heart, Lung, and Blood Institute.
In addition, the National Library of Medicine has issued so-called competitive revision grants geared toward funded scientists "who wish to expand the scope of their research beyond the current funded aims." The deadline for applications is April 21.
With an explicit focus on informatics education, the National Library of Medicine also offers summer informatics research opportunities for high school, undergraduate students, and science educators, including faculty from "non-research intensive institutions."