InstaDeep
Run the Nucleotide Transformer DNA language models via InstaDeep's API for genomics, variant effect, and regulatory sequence prediction.
Overview
InstaDeep is a genomics AI lab and the first-party source of the Nucleotide Transformer, a family of DNA foundation models. If you need hosted access to genomic language models for variant effect prediction, regulatory element detection, or general DNA sequence representation, InstaDeep is the origin lab and a go-to for the DNA and genomics slice of the catalog.
What you can run on InstaDeep
- Nucleotide Transformer — a family of transformer-based DNA language models spanning 500M to 2.5B parameters, pretrained on 3,200+ human genomes and 850 species. The models produce embeddings for downstream genomics tasks such as promoter and enhancer prediction, splice-site detection, chromatin profiling, and variant effect estimation.
This covers the DNA and genomics language modeling domain. The open Nucleotide Transformer weights are published on Hugging Face, and InstaDeep provides hosted API access to run the models.
API access on InstaDeep
InstaDeep offers hosted inference for the Nucleotide Transformer through its API, so genomics teams can generate embeddings and predictions without managing model infrastructure or GPU capacity themselves. This suits computational genomicists, functional-genomics researchers, and bioinformatics teams building on DNA foundation models for regulatory genomics and clinical variant interpretation. The open Nucleotide Transformer weights also remain available on Hugging Face for self-hosted inference and fine-tuning when you prefer to run the models on your own hardware.
Run inference on InstaDeep (1)
DNA foundation models from 500M to 2.5B parameters, trained on 3,200+ human genomes and 850 species for variant effect prediction.