{
  "abstract": "Microbial communities sustain ecosystems and influence host health, yet connecting their molecular diversity to responses under intervention remains difficult. Here we introduce MGM2, a multimodal foundation model integrating DNA language-model representations of microbial sequences, quantitative abundances and sample descriptions. After pretraining on 1.8 million microbiomes, its frozen representations achieved a mean macro-AUROC of 0.91 on later MGnify projects absent from pretraining. These representations also supported fecal microbiota transplantation outcome prediction, recovering 68% of observed high-change taxa. Adding molecular structures enabled prediction of drug effects in nested five-fold evaluation, while measured perturbed states informed subsequent recovery. To interpret the learned representations, sparse decomposition resolved taxonomic, ecological and abundance-related features. Together, these results show how a shared representation of molecular identity and community context connects large-scale microbiome learning to the prediction and interpretation of community change.",
  "abstract_shown": true,
  "affiliations": [
    "Huazhong University of Science and Technology"
  ],
  "article_type": "New results",
  "arxiv": null,
  "author_total": 7,
  "authors": [
    {
      "name": "H. Zhang"
    },
    {
      "name": "Y. Zhang"
    },
    {
      "name": "Y. Qi"
    },
    {
      "name": "T. Liu"
    },
    {
      "name": "L. Zhao"
    },
    {
      "name": "R. Yang"
    },
    {
      "affs": [
        0
      ],
      "name": "K. Ning"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/91fh441bh6j7v.bib",
  "cited_by": 0,
  "doi": "10.64898/2026.07.20.739063",
  "fulltext_url": "https://www.biorxiv.org/content/10.64898/2026.07.20.739063v2.full",
  "id": "91fh441bh6j7v",
  "issns": [],
  "landing_url": "https://www.biorxiv.org/content/10.64898/2026.07.20.739063v2",
  "language": null,
  "licence": "CC-BY-NC",
  "licence_label": "CC BY-NC",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by-nc/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "biorxiv",
      "bioinformatics"
    ],
    [
      "openalex",
      "1312|Gut microbiota and health"
    ]
  ],
  "openalex": "W7170183213",
  "pdf_url": "https://www.biorxiv.org/content/10.64898/2026.07.20.739063v2.full.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-07-20",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/91fh441bh6j7v.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:25:26Z",
      "native_id": "10.64898/2026.07.20.739063",
      "source": "biorxiv",
      "url": "https://www.biorxiv.org/content/10.64898/2026.07.20.739063v2"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "biorxiv:bioinformatics",
      "subfield": "bioinformatics"
    }
  ],
  "title": "MGM2 as a Multimodal Foundation Model for Microbial Community States and Responses",
  "title_text": "MGM2 as a Multimodal Foundation Model for Microbial Community States and Responses",
  "type": "preprint",
  "updated_at": "2026-10-30T12:25:26Z",
  "url": "https://research.draconitic.com/paper/91fh441bh6j7v",
  "venue": "bioRxiv",
  "versions": []
}