{
  "abstract": "Genomic foundation models are revolutionizing biological research and biotechnology, yet their vulnerability to data poisoning remains poorly characterized. Here we show that backdoor attacks can be reliably implanted during fine-tuning across diverse model architectures, tasks and model scales by poisoning only a small fraction of training data with short DNA motifs, including triggers derived from transposon terminal inverted repeats (TIRs). Attack success rates are generally high across encoder-based and autoregressive models, while clean-task performance remains almost unchanged. Increasing model scale does not confer measurable resistance. To detect such threats, we develop a two-stage auditing framework that flags potentially poisoned sequences through localized single-nucleotide mutation sensitivity in model predictions, without prior knowledge of trigger sequences. These results identify downstream fine-tuning as an attack surface for genomic foundation models and motivate stronger data provenance, adversarial evaluation and post-training security auditing.",
  "abstract_shown": true,
  "affiliations": [
    "Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences"
  ],
  "article_type": "New results",
  "arxiv": null,
  "author_total": 12,
  "authors": [
    {
      "name": "S. Ni"
    },
    {
      "name": "Q. Wang"
    },
    {
      "name": "C. Wei"
    },
    {
      "name": "X. Ni"
    },
    {
      "name": "S. Li"
    },
    {
      "name": "Z. Zhao"
    },
    {
      "name": "H. Li"
    },
    {
      "name": "R. Ji"
    },
    {
      "name": "R. Xu"
    },
    {
      "name": "H. Alinejad-Rokny"
    },
    {
      "name": "T. Wang"
    },
    {
      "affs": [
        0
      ],
      "name": "M. Yang"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/frtqcmyxbsq8p.bib",
  "cited_by": 0,
  "doi": "10.64898/2026.07.30.741642",
  "fulltext_url": "https://www.biorxiv.org/content/10.64898/2026.07.30.741642v2.full",
  "id": "frtqcmyxbsq8p",
  "issns": [],
  "landing_url": "https://www.biorxiv.org/content/10.64898/2026.07.30.741642v2",
  "language": null,
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "biorxiv",
      "bioinformatics"
    ],
    [
      "openalex",
      "1702|Adversarial Robustness in Machine Learning"
    ]
  ],
  "openalex": "W7172489999",
  "pdf_url": "https://www.biorxiv.org/content/10.64898/2026.07.30.741642v2.full.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-07-30",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/frtqcmyxbsq8p.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:26:02Z",
      "native_id": "10.64898/2026.07.30.741642",
      "source": "biorxiv",
      "url": "https://www.biorxiv.org/content/10.64898/2026.07.30.741642v2"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "biorxiv:bioinformatics",
      "subfield": "bioinformatics"
    }
  ],
  "title": "Pervasive Backdoor Vulnerabilities in Genomic Foundation Models",
  "title_text": "Pervasive Backdoor Vulnerabilities in Genomic Foundation Models",
  "type": "preprint",
  "updated_at": "2026-10-30T12:26:02Z",
  "url": "https://research.draconitic.com/paper/frtqcmyxbsq8p",
  "venue": "bioRxiv",
  "versions": []
}