Biological models

318 model records in release 2026-10-23-5347695500ed (release manifest, gzip). Showing 193–216; page 9 of 14.

Evaluated configurations, methods and pipelines are listed separately. Names alone do not establish equivalent models or checkpoints.

• MSNovelist

  Metabolomics

  MSNovelist de novo structure generation from fingerprints.

• MULAN

  Proteins and complexes

  Multimodal protein language model that fuses a pretrained sequence encoder with a parameter-efficient adapter for angle-based structure encoding, scored in ProteinGym as a zero-shot substitution baseline from sequence and structure.

• MutBERT

  DNA and genomes

  Masked DNA language model that represents the human genome as allele-frequency distributions from population data, so pretraining concentrates on sites of variation.

• MVP

  Metabolomics

  Contrastive spectrum-to-molecule retrieval model that extends JESTR with multiple views (fingerprint, graph, spectra and consensus spectra); it needs the target's molecular formula.

• NEIMS

  Metabolomics

  Neural electron-ionization mass spectrum predictor, used as a predictor head in FlexMS.

• Neuraldecipher

  Metabolomics

  Neuraldecipher fingerprint-to-molecule decoder.

• NeuralFP

  Protein function

  Neural fingerprint graph network.

• Nicheformer

  Single-cell omics · Spatial omics

  Transformer foundation model pretrained on dissociated and spatial transcriptomics.

• Novae

  Single-cell omics · Spatial omics

  Graph attention foundation model for spatial transcriptomics domains.

• Nucleotide Transformer

  DNA and genomes

  Nucleotide Transformer is a family of DNA encoders pretrained on human or multispecies sequence corpora.

• Nucleotide Transformer v2

  DNA and genomes

  Nucleotide Transformer v2 represents DNA using an encoder pretrained on multiple species.

• Nucleotide Transformer v3

  DNA and genomes

  Nucleotide Transformer v3 genomic language model family (pre- and post-trained checkpoints).

• NuFold

  RNA and transcriptomes · Protein structure

  NuFold: RNA 3D structure prediction method built on the AlphaFold architecture with RNA-specific inputs.

• o1

  Cells and tissues

  OpenAI reasoning language model that generates an internal chain of thought before giving its answer.

• o1-mini

  Cells and tissues

  Faster version of OpenAI's o1 reasoning model, trained with reinforcement learning to think in a chain of thought before answering and particularly effective at coding.

• o1-preview

  Cells and tissues

  Preview release of OpenAI's o1 reasoning model, a large language model trained with reinforcement learning to produce a long chain of thought before answering.

• Omni-DNA

  DNA and genomes

  Autoregressive transformer family pretrained on DNA by next-token prediction and then fine-tuned with added task tokens to solve several genomic tasks, including DNA-to-text and DNA-to-image mapping, in one model.

  Record: gfm2-20261011-datta2026-model-omni-dna

• Omni-DNA

  DNA and genomes

  Family of autoregressive DNA transformers from 20 million to 1 billion parameters, pretrained by next-token prediction and multi-task fine-tuned with added task tokens for classification and cross-modal outputs.

  Record: lit-genomics-20261011-geneb-model-omni-dna

• OmniNA

  DNA and genomes

  Generative nucleotide model pretrained on 91.7 million sequences with their text annotations from many species, fine-tuned to answer sequence tasks in natural language.

• OmniReg-GPT

  Microbes and communities · DNA and genomes

  OmniReg-GPT generative genomic foundation model.

• OpenFold

  Protein structure

  OpenFold is a trainable PyTorch implementation of AlphaFold 2 and AlphaFold-Multimer workflows.

• OpenSpliceAI

  DNA and genomes

  PyTorch reimplementation of SpliceAI retrained on RefSeq MANE v1.3 protein-coding transcripts (GRCh38).

• Optimus 5-Prime

  RNA and transcriptomes

  Convolutional neural network that predicts mean ribosome load from 5' UTR sequence, trained on massively parallel reporter assays of random 5' UTRs and limited by position-specific weights to the lengths seen in training.

• Orca

  DNA and genomes

  Orca 3D genome structure model.

This index reflects a dated catalogue, not an exhaustive census. Source checking does not mean independent reproduction; compare results only under compatible protocols, datasets and metrics.