{"resourceId":"hugging-face-carbon-a-open-gene-annotation-566m-loci-20261008","versions":[{"version":"external-cd1b46d8273c983660efb03f22a3e1b62cb108a33fac8fe453a3825dd20c8568","resource":{"id":"hugging-face-carbon-a-open-gene-annotation-566m-loci-20261008","title":"Hugging Face releases Carbon-A open gene-finding model and a database of 566 million predicted protein-coding loci across 22,617 taxa","organization":"Hugging Face (HuggingFaceBio); wet-lab validation partners ActiveSite and UC San Diego; EMBL-EBI (planned distribution)","sector":"Open AI-for-science models and research data infrastructure (genomics)","geography":"Global; public GenBank/ENA assemblies","publishedAt":"Hugging Face blog, 8 October 2026","publicationDate":"2026-10-08","eventDate":null,"sourceName":"Hugging Face blog","sourceLabel":"Developer release post (self-reported results)","sourceUrl":"https://huggingface.co/blog/HuggingFaceBio/carbon-annotator-genbank-genome-annotation","evidenceClass":"vendor-claim","outcomeClass":"emerging","topics":["infrastructure","governance-procurement","operating-model"],"finding":"Hugging Face released Carbon-A, a 1.2B-parameter model predicting protein-coding regions directly from DNA across eukaryotes, and the Carbon Annotation Database: 48,167 GenBank assemblies, 22,617 taxa, about 27 trillion base pairs and 566 million predicted coding loci (about half the target set; next batch in three weeks; EMBL distribution planned). Self-reported results: macro-averaged nucleotide F1 0.944 on 42 benchmark genomes, ahead of seven named baselines; 0.960 F1 on Tetrahymena; roughly 0.62 complete-CDS support against PacBio Iso-Seq data in four species; confidence-score AUROC 0.876. The post states Iso-Seq shows transcription, not translation or function.","sledRelevance":"Shows open model weights plus a large precomputed prediction database as a research-infrastructure product that university genomics cores and libraries may ingest. Relevant to research computing, data stewardship and validation policy for AI-generated annotations.","evidence":"Opened the Hugging Face blog post via HTTPS (curl, full text extracted and read). Model card also opened; a secondary write-up reports the card warns the evaluation panel was used in training, which was not confirmed verbatim in the extracted card text.","architectureImplications":"Interpretation: institutions can run the model locally on GPUs or consume the hosted database; either way, predictions need provenance (model version, confidence) stored alongside curated annotations.","governanceImplications":"AI-predicted genes should be labelled as candidates with confidence scores in institutional databases, distinct from experimentally supported annotations.","securityPrivacyImplications":"Public non-human genomes; low privacy exposure. Running on sensitive or human data was not described.","caveats":"All benchmark figures are self-reported by the developer; the technical report was not inspected; wet-lab validation covers four well-studied species and shows transcription only; additional wet-lab results are unpublished; one consensus isoform per locus.","streamIds":["research"],"roles":{"sales":"Sales framing: open AI annotation resources are arriving at a scale labs cannot validate themselves. Research computing and genomics cores need a way to ingest, label and verify AI predictions before they flow into publications. A bounded data-governance and validation-workflow engagement fits. This evidence cannot support claims of discovered genes or accuracy beyond the developer's self-reported benchmarks; Iso-Seq support shows transcription, not function.","engineering":"Engineering framing: Carbon-A can be run from the Hub on local GPUs or consumed as a hosted database. Fit depends on GPU access, genome file pipelines and storing per-gene confidence and model version with each prediction. Constraints include self-reported benchmarks and single-isoform output. Validate on one well-annotated in-house genome and compare against curated annotations and local transcript data before wider use.","delivery":"Delivery framing: the work is integrating AI-predicted annotations into a core facility's data without blurring them with curated records. Tasks include provenance tagging, confidence thresholds and a review step. Ownership sits with the genomics core and research data services. Propose acceptance measures such as every imported prediction carrying model version and confidence; risks include over-trusting predictions and drift as new database batches arrive."},"roleDetails":{"sales":{"customerProblem":"Genomics labs and cores face large open AI-predicted annotation sets that they cannot validate individually.","stakeholders":["Genomics core director","Research computing","Research data services/library","PIs in comparative genomics"],"discoveryQuestions":["How are AI-predicted annotations labelled versus curated ones?","What confidence threshold is acceptable for downstream use?","Who validates predictions for in-house organisms?"],"valueHypothesis":"Clear provenance and validation workflows may let labs use open AI annotations faster with less risk of propagating errors; unquantified.","potentialEngagement":"Three-week design of provenance tagging and validation workflow for AI-predicted annotations.","unsupportedClaims":["Confirmed novel genes or protein functions","Independently verified accuracy","Performance on organisms outside the tested set"]},"engineering":{"fit":"University genomics groups with GPU access or need for annotations of poorly studied eukaryotes.","architectureIntegration":"Hub model inference on local GPUs or ingestion of hosted database records into lab data stores with confidence and version fields (interpretation).","prerequisites":["GPU capacity","FASTA/GenBank pipelines","Curated reference genome for comparison"],"constraints":["Self-reported benchmarks","Single consensus isoform","Bacteria and viruses out of scope per secondary reporting"],"security":"Low risk for public genomes; review terms before running on controlled-access data.","proofOfValue":["Compare predictions on one in-house curated genome","Check confidence-score calibration locally","Cross-check novel predictions against local RNA data"]},"delivery":{"work":"Define ingestion, provenance tagging, confidence thresholds and review steps for AI-predicted annotations.","dependencies":["Database batch releases","Local validation data","Core facility staff time"],"ownership":"Genomics core with research data services.","skillsAdoption":"Train users to read confidence scores and distinguish transcription evidence from functional evidence.","governanceCheckpoints":["Provenance schema approval","Threshold sign-off","Re-review when model or database version changes"],"acceptanceCriteria":["Proposed: 100% of imported predictions carry model version and confidence","Proposed: local benchmark run documented before production use","Proposed: candidate genes flagged distinctly in shared databases"],"risks":["Over-trust in predictions","Version drift across batches","Errors propagating into publications"]}},"retrievedAt":"2026-10-11T21:58:00-05:00","enrichedAt":"2026-10-11T21:59:53-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Open weights and data lower barriers for labs studying poorly annotated organisms, but users need training to interpret confidence scores.","procurementImplications":"Open MIT-style release avoids licensing cost but shifts compute and validation costs to institutions.","operatingModelImplications":"Shared, precomputed annotation resources may replace per-lab pipelines; cores need policies for when to trust or re-run predictions.","sourceVerification":{"openedUrl":"https://huggingface.co/blog/HuggingFaceBio/carbon-annotator-genbank-genome-annotation","referenceExcerpt":"Reference annotations and model predictions are both hypotheses about underlying biology","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}