---
title: Biological Language Modeling — CyMetica
url: https://cymetica.com/biological-language-modeling
---

# Biological Language Modeling

DNA (alphabet ACGT), RNA (ACGU) and proteins (20 amino acids) are languages: finite alphabets, words (codons), grammar (regulatory structure) and meaning (function). They can be modeled with the same self-supervised language-model methods used for text, and any other biological data (expression matrices, assays, literature) can be embedded into the same vector space in a standard data-engineering pipeline: ingest, validate, embed, index, join, evaluate.

## Evidence

- Searls DB (2002). The language of genes. Nature 420:211–217. https://doi.org/10.1038/nature01255
- Ji Y et al. (2021). DNABERT. Bioinformatics 37(15):2112–2120. https://doi.org/10.1093/bioinformatics/btab083
- Dalla-Torre H et al. (2024/2025). Nucleotide Transformer. Nature Methods 22:287–297. https://doi.org/10.1038/s41592-024-02523-z
- Nguyen E et al. (2024). Evo. Science 386. https://doi.org/10.1126/science.ado9336
- Chen J et al. (2022). RNA-FM. arXiv:2204.00300. https://arxiv.org/abs/2204.00300
- Rives A et al. (2021). ESM-1b. PNAS 118(15). https://doi.org/10.1073/pnas.2016239118
- Lin Z et al. (2023). ESM-2 / ESMFold. Science 379:1123–1130. https://doi.org/10.1126/science.ade2574
- Zvyagin M et al. (2023). GenSLMs. IJHPCA 37(6):683–705. https://doi.org/10.1177/10943420231201154
- Jumper J et al. (2021). AlphaFold. Nature 596:583–589. https://doi.org/10.1038/s41586-021-03819-2

## Try it

Live sequence similarity over a training set: https://cymetica.com/dna

## Work with us

Delivered by forward deployed engineers (https://cymetica.com/forward-deployed-engineering). Email contact@cymetica.com.