How to Read This Handbook

Published

July 7, 2026

This handbook is organised by research object and workflow: foundations, molecular discovery and design, cells, tissues, systems biology, organismal and environmental biology, therapeutic discovery and translation, and research systems, practice, and governance. Each major chapter starts with the same field-reference questions, then uses the same evidence-utility questions so a reader can locate the evidence level before acting on it. The handbook is a reference, not a tutorial; you do not need to read it cover to cover.

Fast Path by Topic

The shortest paths to handbook value, by topic:

Reading Paths by Role

The audience is varied; the question is shared. When does an AI output deserve experimental attention? The shortest path through the handbook depends on your role.

Computational biologist

You need capability tier for each model class, known failure modes, and how to design benchmarks that reflect your actual question. Start with AI for the Life Sciences for orientation, then Evaluation Principles for the discipline, then the model classes most relevant to your work. Pay close attention to the critical-evaluation sections in each chapter; single-cell baseline critiques and docking validity checks change how much weight to give headline numbers.

Biotechnology team lead

You need build-vs-buy framing, license diligence, and what the open-source alternatives are when frontier releases are restricted. Read the Executive Summary, then the chapters on the model classes your program depends on, then Workforce, Compute, and Institutional Readiness for team composition. Use the AlphaFold 3 restricted-release and Boltz/Chai open-source response as the licensing-risk case study.

Drug discovery scientist

You need to know where AI shifts a stage gate and where it does not. Read Target Identification, Small Molecule Generation and ADMET, Clinical Trial AI, and Translational Evidence and Failure Modes in sequence. The Insilico Medicine ISM001-055 case (Ren et al., 2025) is the furthest-advanced AI-discovered small molecule in clinical trials; the chapter discusses what Phase IIa evidence means and what it does not.

Physician-scientist

You need translation between bench AI and clinical decision-making. Read Variant Effect Prediction for ACMG/AMP framework integration, Clinical Trial AI for regulatory framing, and Translational Evidence and Failure Modes. Protein Structure Prediction is relevant if your work touches drug design; Single-Cell Foundation Models is relevant if your work uses single-cell data.

Synthetic biologist

You need design-tool selection, automation integration, and dual-use awareness. Read Protein Design and Engineering, Synthetic Biology Design Tools, Self-Driving Laboratories, and Information Hazards in Capability Research. The screening discipline for DNA synthesis orders is essential for any work that uses AI sequence generation.

Plant, ecology, or environmental researcher

You need a structure that treats plant and ecological systems as core life sciences, not as side cases. Read Plant, Crop, and Agricultural AI or Environmental and Ecological AI, then Systems Biology and Multiscale Modeling, then Evaluation Principles. The critical question is measurement validity: whether the model is learning biology, sampling effort, geography, season, or platform artifacts.

Neuroscientist or aging researcher

You need to separate biological representation from intervention evidence. Read Neuroscience AI and Brain Foundation Models or Aging and Longevity Biology AI, then Virtual Organisms and Digital Biology, then Benchmarks for Bio AI. Neural decoding, brain foundation models, aging clocks, and geroscience models can be useful research instruments without proving that a model understands cognition, healthspan, or organism-level causality.

Graduate student or early-career researcher

You need conceptual entry points and canonical citations. Read History of AI in the Life Sciences for the five-decade arc, then AI for the Life Sciences for the scope framing, then Evaluation Principles for the framework that travels across subdomains. Pick a model-class chapter for the area you work in and use its citations as a starting reading list.

Research program leader

You need funding and program-prioritization framing, which capabilities are infrastructure-grade vs. research-grade, and how to evaluate proposals that invoke AI. Read the Executive Summary, AI for the Life Sciences, Workforce, Compute, and Institutional Readiness, and the chapter for any capability claim a proposal rests on. The five-question standard is the right lens for distinguishing program-eligible claims from research aspirations.

The Chapter Reference Standard

Used throughout the handbook. Every major chapter begins by answering the same five reference questions:

  • What is this field trying to solve? The biological or research problem that defines the area.
  • What is the core idea? The assumptions, constraints, and terms that prevent misreading the field.
  • What is the current state of the field? The active methods, model classes, and evidence landscape.
  • What do we know, and what remains open? The settled reference points, unresolved questions, and limits of current evidence.
  • Why does this matter? The scientific, translational, or program implication.

The Chapter Utility Standard

Used throughout the handbook. Every major chapter answers the same five questions:

  • What is demonstrated? Supported by published evidence in peer-reviewed venues, official documentation, or reproducible blind benchmark results. The evidence must be specific to a defined task and dataset.
  • What is theoretical? Plausible given current methods but not yet established for routine use. The capability has been shown in selected systems or narrow tasks without proven generalisation.
  • What is beyond current capability? Not supported by credible evidence with current systems. The capability is aspirational, has been demonstrated only in toy settings that do not generalise, or requires evidence that has not been produced.
  • What would make this more promising? The benchmark, prospective experiment, wet-lab confirmation, clinical endpoint, ecological field test, or independent reproduction that would make the area stronger or reveal its limits.
  • What should researchers, biotech teams, funders, and program leaders do with this? The practical consequence for evaluation, adoption, funding, experiment design, or deferral.

If a claim cannot answer these questions, the claim is not specific enough. The point is not to be conservative; it is to be precise. In practice, the evidence standard borrows from blinded community benchmarks such as CASP (Kryshtafovych et al., 2024), biology-specific reporting frameworks such as DOME (Walsh et al., 2021), and validity checks such as PoseBusters (Buttenschoen et al., 2024).

Citation Conventions

Citations are inline hyperlinks. The format depends on source type:

  • Peer-reviewed papers: (Author et al., Year) linking to the DOI. Example: (Jumper et al., 2021) linked to the AlphaFold 2 Nature paper.
  • Preprints: (Author et al., Year, preprint) linking to the bioRxiv, medRxiv, or arXiv version. Example: (Zambaldi et al., 2024, preprint) linked to AlphaProteo on arXiv.
  • Company sources: (Company, Month Year) linking to the official source. Example: (OpenAI, June 2024) linked to the Color Health partnership announcement.
  • Government and program pages: (Org, program page) or (Org, Year) linking to the official page. Example: (ARPA-H IGoR, 2026).

Every citation in this handbook is verified against Crossref, PubMed, publisher records, or official source pages before being written. Hallucinated DOIs and author-attribution errors are caught and corrected; if you spot one, please report it.

How to Triage a Capability Claim

The discipline for any new claim (paper, preprint, vendor pitch, conference talk):

  1. Name the biological object. Sequence, structure, ligand, cell state, tissue, experiment, or clinical endpoint.
  2. Name the task. Prediction, generation, ranking, classification, or planning.
  3. Identify the evidence. Peer-reviewed benchmark? Prospective experiment? Self-reported? Independent reproduction?
  4. Place the claim in a utility question. What is demonstrated, what is theoretical, or what is beyond current capability?
  5. Name the falsifying experiment. What measurement would change your assessment?

A claim that survives these five steps is worth evaluating. A claim that fails any of them is not yet ready to inform a program decision.

Cross-Handbook Navigation

This handbook sits in a series. For the clinical, public health, and biosecurity layers of AI in biology, see the Companion Handbooks section on the welcome page. The Life Sciences AI Handbook focuses on the discovery layer; the others focus on their respective downstream layers. Each is self-contained.

A Note on Reading Strategy

The handbook is a reference, not a linear textbook. Most readers should start with the Executive Summary, use the role-specific paths above, then return to individual chapters when a model class, evidence standard, or institutional decision needs review. The Quick Reference appendix consolidates the chapter summaries for scanning before a deeper read.