Skip to content
CAAIL

CAAIL · Cellular Agriculture AI Library

Cell Ag × AI

A curated library at the intersection of cellular agriculture and AI, and a grounding database an AI assistant can query, so research and industry questions can be answered from a fixed, citable index.

CAAIL is in beta. We're re-checking earlier classifications and reorganizing the research areas for v2. See the roadmap Tell us what you think

Indexing 348 references, 143 software tools, 150 databases, 218 datasets.

Counted at build time

Point an AI agent at CAAIL

No download, no command line. In Claude Science:

  1. Settings → Skills → Add skill → Import from GitHub
  2. tucca-cellag/caail
  3. Preview → Import

Claude Science

Beta, macOS or Linux, Pro plan or above. Imports the query and contribute skills, untick either; pinned to that commit, and Check for updates moves them forward. Claude Science skills docs →

Dedicated ChatGPT and Gemini integrations are in progress. Until then the Any agent prompt works in both; it just does not persist.

Explore the library

Each collection is published as Markdown on GitHub.

Start here

New to cell ag?

Watch the field foundations and pick up the reference reading.

New to AI?

Start with ML learning playlists built for bench scientists.

Looking for a method?

Filter the Papers matrix by AI method or research area.

Need training data?

Browse datasets by species, or benchmarks for model evaluation.

Recently added

2026-09-22 Paper EFSA allergenicity opinion and Swartz GF cost report
2026-09-02 Paper iUmami-SCM and CRISPR-GPT each placed in one more matrix cell
2026-08-25 Paper The Hybrid Mechanistic-ML and Comparative Studies rows
2026-08-20 Paper Metabolic Modeling and Food Safety Prediction columns
2026-08-20 Software Three allergenicity predictors
View the full changelog →

Topics

Is your own area in here, and how thinly?

8 subject themes cut across papers, software, databases and dataset entries. The counts are uneven, and they count what CAAIL holds, not how much work each area has produced. Tags were set by rule or by hand, and they have not been audited end to end. Some known errors are not yet fixed, so if your own area looks wrong here, please tell us.

16 finer tags sit beneath these 8 themes. A tag is added only where several items share it. Open the topic hub →

Why CAAIL

The same question, answered from memory and from an index

CAAIL is a grounding database: a static index an AI agent queries directly, so an answer can be assembled from the corpus instead of recalled from training. It holds 230 primary-research papers, each placed in at least one cell of a 27 × 8 matrix of AI method against research area, alongside 76 reviews and perspectives, 42 reference-work entries, 143 software tools, 150 databases and 218 dataset entries. Every figure in this section is read from that corpus when the page is built.

Answered from model memory

You

Which papers applied deep learning to cell-line engineering for cultivated meat, and which of them released code I can run?

AI

Convolutional and transformer models have both been used to guide cell-line engineering, for example Zhang et al. (2021) in Metab. Eng.… [nothing here tells you whether this exists]

  • A citation is generated, not retrievedThe reference is produced token by token, so a well-formed DOI is not evidence that the work exists
  • Ask twice, get two answersDecoding samples from a distribution, so a pinned model on the same day need not agree with itself
  • There is no result for "found none"A gap in the training data and a gap in the field produce the same reply
  • Frequency stands in for relevanceWhat is heavily cited and plainly titled is what is recalled; a tool with an obscure name is not
  • Reuse terms do not travel with the answerA tool arrives named but unlicensed, so nothing separates one you can ship from one whose terms your product would inherit

Answered from CAAIL

You

Which papers applied deep learning to cell-line engineering for cultivated meat, and which of them released code I can run?

API

matrix.json → Deep Learning × Cellular Engineering

AI

6 papers: Li 2020 (#5), Magnusson 2024 (#122), Adduri 2025 (#57) and 3 more. 3 link runnable code. [ids resolve, corpus dated 2026-11-14]

  • All 216 method × area cells are enumerated138 hold no indexed paper, and each is returned as an empty result rather than skipped
  • An empty cell is returned with its own limit attachedCAAIL is a curated selection, so an empty cell is a limit of the index, not evidence that no such work exists
  • Entries are chosen by review, not by a popularity signalAn obscure tool and a well-known one are listed on the same basis, because inclusion is a recorded judgement
  • One subject axis across papers, software, databases and dataset entriesA design question returns them together, not one type at a time
  • A licence tier per tool, and a stated gap where there is noneOf 143 software tools, 79 are permissive and 22 copyleft, so a derivative inherits the terms; 9 are restricted and 33 record none
  • The same query returns the same rowsA fixed corpus dated 2026-11-14, so a result can be quoted and re-checked by someone else

The left exchange is constructed and its citation is invented: it illustrates the failure, it does not report one. The right exchange is generated when this page is built, from the same matrix.json an agent fetches.

Every reference id above resolves to an entry with its DOI. Browse the matrix →

What you can ask

Ask something specific

Four questions a working researcher might ask, run against the endpoints an agent would query.

These answers are built from what the endpoints return, before an agent reasons further over them. The counts are computed from the corpus when this page is built; other figures come from the paper, guideline or resource each card names.

Experiment design × existing deposits

I am planning an RNA-seq timecourse on primary bovine satellite cells through differentiation, budget for about 12 libraries. I would like this to be integrable with existing data rather than standing alone. Which public datasets could mine realistically be combined with, and how were they designed?

Answered from datasets.json

A bulk 0 to 96 h timecourse is public under serum starvation. In media described as serum-free, CAAIL's bovine inventory holds bulk RNA-seq only at 0 h and 72 h

  • Public at the full gridGSE173198 is bulk RNA-seq at 0/24/48/72/96 h in biological quadruplicates, with differentiation induced by serum starvation
  • Serum-free differentiationGSE173196 is bulk, sampled only at 0 h and 72 h in biological triplicates, and compares serum-free differentiation media with 2% FBS. GSE240556 is single-nucleus RNA-seq covering all five timepoints in serum-free differentiation medium
  • Regulatory-element annotationThe bovine FAANG epigenome atlas's processed annotations are open on Zenodo, but the accession IDs for its newly generated reads are withheld until manuscript acceptance

GSE173198 and GSE173196 are subseries of one GEO SuperSeries, GSE173199. CAAIL holds 45 bovine datasets as a curated selection, so a series missing here may still exist elsewhere.

Media optimization × metabolic modeling

We are trying to cut the cost of our serum-free medium and have reached the point where one-factor-at-a-time screening has stopped paying. I have a draft genome-scale model for our species but no experience closing the loop between a model and a screen. What work couples a metabolic model to an experimental design loop, and what did those groups actually have to measure to make the model useful?

Answered from matrix.jsonpapers.json

13 papers here are genome-scale or metabolic-network work. The matrix reaches 3 of them

  • Bayesian Optimization × Media Optimization8 papers. 4 link released code, and 1 of the 8 links both code and a data deposit
  • The other genome-scale workThe other 10 carry no matrix method, because they are filed as reference works and reviews, not as method-to-area research, so a search of the matrix does not return them
  • A model-guided media designGomez Romero 2026 (#240), "iSsus3744: A genome-scale model-guided strategy for rational media design for cultivated pork", constrains its model with biomass composition measured in pig muscle and fat tissue and with uptake and excretion fluxes measured in a Duroc muscle satellite-cell line, then validates it by amino-acid supplementation. It is a preprint and has not been peer reviewed

What each group measured is in its methods section, not in any metadata field, so this answer ends in papers to read. It shows where the coupling is indexed in CAAIL, not everyone who has done it.

Allergenicity × novel-food dossier

We have engineered a bovine line expressing a couple of non-native proteins and we are scoping a novel-food submission. Before I commission wet-lab allergenicity work, what can I screen the sequences against computationally, and has anyone in cultivated meat already been through this whose submission I can read?

Answered from catalog.json

Allermatch automates the FAO/WHO and Codex sequence comparison. The catalogue lists 9 other allergenicity tools beside it

  • The sequence comparisonAllermatch automates the comparison recommended by the 2001 FAO/WHO expert consultation and the Codex guideline. It slides an 80-amino-acid window with a default threshold of 35% identity, and runs an exact-match search with a default length of 6 amino acids
  • Other allergenicity toolsAllerCatPro 2.0, AllerTOP v2, AllergenFP, AlgPred 2.0, ALLERDET, AllergenAI, AllerTrans, NetAllergen, ChAlPred. Their catalogue entries say what each screens and how
  • Allergen reference databasesWHO/IUIS Allergen Nomenclature Database, AllergenOnline (FARRP), COMPARE Database, SDAP 2.0 (Structural Database of Allergenic Proteins), Allergome, AllFam (Database of Allergen Families)
  • Published precedentsFDA Inventory of Completed Pre-market Consultations for Human Food Made with Cultured Animal Cells publishes, per product, the file number, the sponsor's own safety submission and the agency's response

None of these screens is the full assessment. The allergenicity annexes of the Codex guidelines for foods from recombinant-DNA plants and animals (CAC/GL 45-2003, CAC/GL 68-2008) weigh the protein's source, its sequence similarity to known allergens and its structural properties, such as resistance to digestion and to heat, since "no single criterion is sufficiently predictive". They call for testing against sera, where available, when the source is known to be allergenic or the sequence shares homology with a known allergen. Check each resource's scope and update date at its source before relying on it.

Sensory prediction × small samples

We run a small cultured-pork line and have GC-MS volatile panels plus trained-panel scores for about 40 samples, nowhere near enough to train on. Before I commit to another year of panel sessions, which approaches have actually predicted sensory outcomes from instrumental measurements in meat, and what did those groups need in sample counts and reference standards to make it work?

Answered from matrix.jsonpapers.json

Classical methods outnumber deep learning in the sensory column, and 11 of its 52 papers release code

  • Methods in the column32 papers across the ensemble, SVM, chemometrics and nearest-neighbour rows, against 20 on the deep learning, CNN, GNN, GAN / VAE and foundation-model rows combined. The largest cell in the column is Ensemble Learning × Sensory Prediction, at 23 papers
  • Sample countsColantonio 2022 (#72) ran a subsampling analysis in tomato: training sets from 50 to 170 with 39 held out, on 209 samples. Accuracy mostly rose across that range, though it was already relatively high for some traits at 50
  • Code releaseSensory papers release code at 11 of 52 (21%), against 83 of 178 (47%) across the rest of the primary literature. Every sensory reference is primary research, so the two groups are comparable

What each group measured is in its methods section.

Open the Papers Explorer → betaCorpus dated 2026-11-14

Connect your agent

CAAIL is static JSON on GitHub, so there is nothing to host and nothing to authenticate. The paths differ in how long the setup lasts, and in what each one asks of you.

No download, no command line. In Claude Science:

  1. Settings → Skills → Add skill → Import from GitHub
  2. tucca-cellag/caail
  3. Preview → Import

Claude Science

Beta, macOS or Linux, Pro plan or above. Imports the query and contribute skills, untick either; pinned to that commit, and Check for updates moves them forward. Claude Science skills docs →

Dedicated ChatGPT and Gemini integrations are in progress. Until then the Any agent prompt works in both; it just does not persist.

Linked external resources are independent of TUCCA and Tufts University and remain under their own licenses.

Last updated: