Decoding the Molecular Universe -- Workshop Report

Thomas Metz, Joshua Adkins, P. B. Armentrout, Patrick Chain, Fanny Chu, Courtney D. Corley, John Cort, Elizabeth Denis, D W Drell, Katherine Duncan, Robert G. Ewing, Facundo M. Fernández, Oliver Fiehn, Neha Garg, Stefan Grimme, Christopher S. Henry, Robert L. Hettich, Tobias Kind, Roger G. Linington, Gary W. Miller, Trent R. Northen, Kirsten E. Overdahl, Ari Patrinos, Daniel Raftery, Paul Rigor, Richard Smith, Jon R. Sobus, Justin Teeguarden, Ákos Vértes, Katrina M. Waters, Bobbie‐Jo Webb‐Robertson, Antony Williams, David S. WishartarXiv 2023 · DOI · arXiv

TL;DR

This workshop report outlines a grand-scale vision to "decode the molecular universe" by identifying and quantifying every small molecule (<1500 Daltons) in biological and environmental systems. Modeled after the success of the Human Genome Project, the proposed "Chemistry of Life Project" aims to move beyond current analytical bottlenecks using a combination of "hyper-orthogonal" instrumentation, AI-driven in silico predictions, and massive data standardization.

What Problem Does This Paper Address?

While we have successfully sequenced the human genome and made massive strides in proteomics, the "molecular universe"—the millions of small molecules like metabolites, signaling compounds, and environmental toxins—remains largely a mystery.

The authors argue that the fundamental analytical paradigms for measuring these molecules haven't changed significantly in 50 years. We currently rely on matching experimental data against physical chemical standards. However, while there are roughly 2.1 million "known or knowable" chemicals, experts estimate there are at least 25 million unknown compounds that influence biology. We simply cannot synthesize enough standards to identify them all. This "dark metabolome" represents a massive gap in our understanding of disease, agriculture, and national security.

Methods

As a workshop report, the "methods" described are the strategic pillars of the proposed Chemistry of Life Project. The report advocates for a shift from "concept-driven" research to a "tool-driven" infrastructure project. Key proposed approaches include:

  • Hyper-orthogonal Instrumentation: Developing the "Frankentrometer"—an idealized, standalone instrument that integrates multiple orthogonal measurements (e.g., mass spectrometry coupled with ion mobility and spectroscopy) to identify molecules without needing reference standards.
  • In Silico Annotation: Leveraging AI/ML to predict molecular properties (like tandem mass spectra or collision cross-sections) with 10–20% higher accuracy than current models to enable "reference-free" identification.
  • Massive Parallelization: Scaling conventional fractionation and isolation methods (NMR, X-ray crystallography) through automation to build a foundational library of molecular scaffolds.
  • Standardized Data Frameworks: Establishing a "well-engineered" system for metadata and data provenance to ensure that data generated today remains reusable for future AI training.

Datasets

The report references several existing chemical and metabolomic repositories that serve as the starting point for this effort:

  • Public Databases: PubChem, HMDB (Human Metabolome Database), COCONUT (Natural Products), and the EPA’s DSSTox.
  • Data Repositories: The Metabolomics Workbench and the NORMAN Suspect List Exchange.
  • Proposed New Benchmark: A curated library of "molecular scaffolds" representative of all small molecule chemistries found in nature, intended to serve as a physical and digital resource for benchmarking new technologies.

Key Insights and Ideas

  • The Investment Gap: The NIH Common Fund Metabolomics Program (the largest U.S. effort to date) received only ~6.7% of the investment allocated to the Human Genome Project (HGP), despite small molecules being chemically more complex than DNA.
  • Tool-Driven Revolutions: Quoting Freeman Dyson, the report notes that science is often more successful at "discovering new things than explaining old ones." The HGP succeeded because it was an engineering/infrastructure project, not just a series of hypothesis-driven experiments.
  • The "Frankentrometer": A provocative concept for a next-gen device that would operate without reliance on chemical libraries, potentially miniaturized for field use in national security or environmental monitoring.
  • The Chemical Tapestry of America: A proposed sub-target to map the small molecule composition of every square mile of the country.

Key Results

The workshop culminated in the definition of The Chemistry of Life Project, structured into four primary thrusts:

  1. Thrust 1: Create a systematic framework for data collection and processing to enable seamless integration of current and future data.
  2. Thrust 2: Assemble a curated library of molecular scaffolds representative of nature's chemistry to train predictive AI models.
  3. Thrust 3: Develop next-gen experimental and computational tools and apply them to "minimally complex" organisms (e.g., bacteria or insects).
  4. Thrust 4: Determine the complete small molecule composition of a composite human reference sample.

Limitations and Future Work

The authors acknowledge several significant hurdles:

  • Funding and Policy: Unlike the HGP, which had strong bipartisan political support, a "molecular universe" project currently lacks a unified "hook" to excite the public and policymakers.
  • Instrument Development: There has been a relative reduction in funding for instrument development at universities, which is critical for creating the "hyper-orthogonal" tools needed.
  • Complexity: Small molecules do not follow the linear, predictable logic of the genetic code, making in silico prediction far more difficult than DNA sequencing.
  • Next Steps: The "founding coalition" of experts plans to continue refining the concept and seeks formal studies by the National Academies to move the project from an idea to a national policy.

Summary updated

Abstract

On August 9-10, 2023, a workshop was convened at the Pacific Northwest National Laboratory (PNNL) in Richland, WA that brought together a group of internationally recognized experts in metabolomics, natural products discovery, chemical ecology, chemical and biological threat assessment, cheminformatics, computational chemistry, cloud computing, artificial intelligence, and novel technology development. These experts were invited to assess the value and feasibility of a grand-scale project to create new technologies that would allow the identification and quantification of all small molecules, or to decode the molecular universe. The Decoding the Molecular Universe project would extend and complement the success of the Human Genome Project by developing new capabilities and technologies to measure small molecules (defined as non-protein, non-polymer molecules less than 1500 Daltons) of any origin and generated in biological systems or produced abiotically. Workshop attendees 1) explored what new understanding of biological and environmental systems could be revealed through the lens of small molecules; 2) characterized the similarities in current needs and technical challenges between each science or mission area for unambiguous and comprehensive determination of the composition and quantities of small molecules of any sample; 3) determined the extent to which technologies or methods currently exist for unambiguously and comprehensively determining the small molecule composition of any sample and in a reasonable time; and 4) identified the attributes of the ideal technology or approach for universal small molecule measurement and identification. The workshop concluded with a discussion of how a project of this scale could be undertaken, possible thrusts for the project, early proof-of-principle applications, and similar efforts upon which the project could be modeled.
BibTeX
@misc{metz2023decoding,
  author = {Thomas Metz and Joshua Adkins and P. B. Armentrout and Patrick Chain and Fanny Chu and Courtney D. Corley and John Cort and Elizabeth Denis and D W Drell and Katherine Duncan and Robert G. Ewing and Facundo M. Fernández and Oliver Fiehn and Neha Garg and Stefan Grimme and Christopher S. Henry and Robert L. Hettich and Tobias Kind and Roger G. Linington and Gary W. Miller and Trent R. Northen and Kirsten E. Overdahl and Ari Patrinos and Daniel Raftery and Paul Rigor and Richard Smith and Jon R. Sobus and Justin Teeguarden and Ákos Vértes and Katrina M. Waters and Bobbie‐Jo Webb‐Robertson and Antony Williams and David S. Wishart},
  title = {{Decoding the Molecular Universe -- Workshop Report}},
  year = {2023},
  eprint = {2311.11437},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2311.11437},
}
Figure