Description:
Fast-Track proposals will not be accepted.
Number of anticipated awards: 1-2 Budget (total costs, per award): Phase I: up to $225,000 for 6-12 months.
Summary
NIAAA supported studies in genomics, imaging, electrophysiology and optogenetics, electronic health records, and personal wearable devices presents new challenges in analyses and interpretations and opportunities for discovery. The NIAAA data sharing policy (NOT-AA-18-101), effective January 25, 2019, expects that investigators and their institutions will submit their grant-related human subjects data to an NIAAA-sponsored data repository. The future data sources will be combined with significant current data repositories and archives, including: database of Genotypes and Phenotype, dbGaP, https://www.ncbi.nlm.nih.gov/gap, Collaborative Studies on Genetics of Alcoholism (COGA), https://niaaagenetics.org/, National Epidemiologic Survey on Alcohol and Related Conditions-III (NESARC-III), and the NIMH Data Archive, NDA, https://data-archive.nimh.nih.gov/ to form a large and rich source for analysis of alcohol use, suitable for analysis by data science methods. Examples of electronic health records (EHRs) include: (NIH-funded) Health Care Systems Research Collaboratory, https://commonfund.nih.gov/hcscollaboratory, (NIH-funded) Clinical and Translational Science Awards (CTSA) Program, https://ncats.nih.gov/ctsa/about/hubs, Health Care Systems Research Network, http://www.hcsrn.org/en/.
Data science includes and extends beyond bioinformatics and computational neuroscience to discover new relationships and pathways for complex systems of normal human function and during adaptations due to disorders or disease. However, many of the tools needed to answer questions in alcohol research require specific applications, algorithms or toolkits that are not currently available. User-friendly methods and applications program interfaces (APIs) for retrieving metadata and data from the repositories for secondary analyses are not currently available. Alcohol researchers require assistance from data scientists to appreciate the power of tools such as machine learning, deep learning and artificial intelligence and the skills for programmers to implement analysis methods to answer key questions about alcohol use.
NIAAA is interested in analytical approaches and tools that can integrate data (i.e. genetic, social, economic, EHR, treatment approaches) to predict the development of alcohol use disorder, or the effectiveness of interventions that reduce or delay the onset or progression of alcohol use disorder, or guide effective treatment and management strategies for alcohol use disorder, including recovery and relapse. One possible example is a tool that assists researchers in developing a risk algorithm for alcohol use disorder when researchers combine multiple existing data sets. The tool could parse out age, sex, race/ethnicity, with alcohol use measures (quantity/frequency/binge episodes), consequences and other risk factors.
Project Goal
The goal is to develop data science analysis algorithms, mathematical models, and software tools for use in alcohol research, integrating data across disciplines and clinical and basic sciences realms.
Phase I Activities and Deliverables
Specific deliverables may include any of the following:
• New algorithms for integrative analysis of current NIAAA and public ‘big data’ sets, including machine learning, deep learning, artificial intelligence, data mining and other model based and model-free approaches.
• Software applications for data interfaces for aggregation, imputation, harmonization, or visualization of data from multiple sources, including current and future NIH data systems (i.e. NCBI (National Center for Biotechnology Information), dbGaP (database of Genotypes and Phenotypes), National Institute of Mental Health Data Archive), or other studies of alcohol research.
• Algorithms and/or software tools for improving data collection, i.e. smart phone apps, extraction of specific alcohol research parameters from existing large databases and established public health studies, biological sensors or wearable devices, natural language processing for analysis of survey data.
• Generation and validation of computational and/or systems biology models of alcohol exposure and use on cellular, organ, network, or organism scales. Multiscale models are appropriate, along with models that include data from clinical and basic science research.
Activities and deliverables are expected to use currently available data sets and databases. Offerors should discuss potential deliverables with NIAAA-supported researchers to determine research needs and goals. All funded NIAAA studies can be found in the public database, NIH RePORTER, https://projectreporter.nih.gov/reporter.cfm. The generation of new primary data is not supported by this topic.
