This RFI seeks input from individuals and stakeholders throughout the scientific research community and the public regarding any of the following topics:

1. Comments or caveats on disparate data type or format collection (e.g., wearable device data, surveys, electronic health records, -omic and dietary data, diagnostic imaging data) and needs as to how they could be made ‘AI-ready’ (e.g., multifactoriality, well-annotated metadata, sufficient sample size, clear data provenance, etc.) – assuming constant data provenance and quality from data collection and measurement. Comments as to curation needs or requirements for data (e.g., multi-modal and multi-fidelity) collected to make them integrable and ‘AI-ready’ to enable modeling to be actionable. Comments as to important provenance, privacy, and ethical considerations associated with data collected from this program.
2. Consideration of computational and modeling approach challenges, as well as important computational and technical parameters needed to developing algorithms for predicting precision diet recommendations (e.g., when developing algorithms, deciphering scientific discovery, identifying disease-risk biomarkers, or improving accuracy from integrated multi-modal/-parameter data sets). Comments as to collaborative tools for visualization of multiscale (e.g., temporal, biological scale) and multi-modal data as well as other analyses to aid research or implementation of discovery resulting from the initiative.
3. Computational, analytical, system science or modeling resources or tools which NIH should consider adding to the All of Us Researcher Workbench to leverage the data sets that will be generated by this study. In addition to the genetics, survey and electronic health records that All of Us already collects, new data sets are expected to include metagenomics, transcriptomics, metatranscriptomics, metabolomics, dietary information, meal chanllenge data, food images, health disparities, social and behavioral determinants of health, etc. [Respondents may suggest open source (e.g., AMON or QIIME2, DeepVariant) or commercial resources/tools].
4. Opportunities for the NIH to partner in achieving the goals of the Nutrition for Precision Health program with dot.org-s, dot.com-s or dot.edu-s (e.g., access to existing accessible data sets, platform developers). Comments on challenges and opportunities for engaging and collaborating with AI and machine learning researchers from math and engineering fields.
5. Any other topic the respondent feels is relevant for the NIH to consider in developing this strategic plan.

Response Date: Nov. 15, 2020