{
  "abstract": "Generative machine learning models offer a powerful framework for therapeutic design, by efficiently exploring large spaces of biological sequences enriched for desirable properties. Unlike supervised learning methods, which require both positive and negative labeled data, generative models such as LSTMs can be trained solely on positively labeled sequences, for example, high-affinity antibodies. This is particularly advantageous in biological settings where negative data are scarce, unreliable, or biologically ill-defined. However, the lack of attribution methods for generative models has hindered the ability to extract interpretable biological insights from such models. To address this gap, we developed Generative Attribution Metric Analysis (GAMA), an attribution method for autoregressive generative models based on Integrated Gradients. We assessed GAMA using synthetic datasets with known ground truths to characterize its statistical behavior and validate its ability to recover biologically relevant features. We further demonstrated the utility of GAMA by applying it to experimental antibody-antigen binding data. GAMA enables model interpretability and the validation of generative sequence design strategies without the need for negative training data. Author summary: We have developed a new way to look inside AI language models that create biological sequences such as antibodies, even when only examples of successful binders are available. Traditional design methods need both positive and negative examples, but in many biomedical settings negative data are scarce, unreliable, or undefined. By training a model on only the desirable (positive) sequences and then comparing the model’s internal “gradient” signals to those of an untrained reference, we can identify which positions in a protein sequence the model deems important for binding. This approach, which we call Generative Attribution Metric Analysis (GAMA), lets us recover known binding motifs from synthetic data, confirm biologically relevant positions in simulated antibody‑antigen interactions, and highlight key residues in real experimental antibody datasets. The method therefore adds a layer of interpretability to otherwise “black-box” AI-based protein design tools, helping researchers understand why a generated sequence is predicted to work and enabling more confident use of these models in therapeutic development and other areas where only positive examples exist.",
  "abstract_shown": true,
  "affiliations": [
    "Department of Immunology and Oslo University Hospital, University of Oslo, Oslo, Norway",
    "Institute for Machine Learning, Johannes Kepler University Linz, Linz, Austria",
    "Department of Informatics, University of Oslo, Oslo, Norway",
    "Translational Immunology, Department of Biomedicine, University of Basel, Basel, Switzerland",
    "Imprint Labs, LLC, New York, New York, United States of America"
  ],
  "article_type": "Research Article",
  "arxiv": null,
  "author_total": 7,
  "authors": [
    {
      "affs": [
        0
      ],
      "name": "Robert Frank"
    },
    {
      "affs": [
        1
      ],
      "name": "Michael Widrich"
    },
    {
      "affs": [
        2
      ],
      "name": "Rahmad Akbar"
    },
    {
      "affs": [
        3
      ],
      "name": "Günter Klambauer"
    },
    {
      "affs": [
        4
      ],
      "name": "Geir Kjetil Sandve"
    },
    {
      "name": "Philippe A. Robert"
    },
    {
      "name": "Victor Greiff"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/3pwtdp3z2h5a7.bib",
  "cited_by": 0,
  "doi": "10.1371/journal.pcbi.1014805",
  "fulltext_url": "https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014805",
  "id": "3pwtdp3z2h5a7",
  "issns": [
    "1553-7358"
  ],
  "landing_url": "https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014805",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "journal",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "openalex",
      "1702|Explainable Artificial Intelligence (XAI)"
    ],
    [
      "plos",
      "/Biology and life sciences/Biochemistry/Proteins/Immune system proteins/Antigens"
    ],
    [
      "plos",
      "/Biology and life sciences/Immunology/Immune system proteins/Antigens"
    ],
    [
      "plos",
      "/Biology and life sciences/Molecular biology/Molecular biology techniques/Sequencing techniques/Protein sequencing"
    ],
    [
      "plos",
      "/Biology and life sciences/Neuroscience/Cognitive science/Cognitive psychology/Language"
    ],
    [
      "plos",
      "/Biology and life sciences/Physiology/Immune physiology/Antigens"
    ],
    [
      "plos",
      "/Biology and life sciences/Psychology/Cognitive psychology/Language"
    ],
    [
      "plos",
      "/Computer and information sciences/Artificial intelligence/Machine learning"
    ],
    [
      "plos",
      "/Computer and information sciences/Data management/Data visualization"
    ],
    [
      "plos",
      "/Engineering and technology/Signal processing/Signal to noise ratio"
    ],
    [
      "plos",
      "/Medicine and health sciences/Immunology/Immune system proteins/Antigens"
    ],
    [
      "plos",
      "/Research and analysis methods/Chemical characterization/Binding analysis"
    ],
    [
      "plos",
      "/Research and analysis methods/Database and informatics methods/Bioinformatics/Sequence analysis/Sequence motif analysis"
    ],
    [
      "plos",
      "/Research and analysis methods/Molecular biology techniques/Sequencing techniques/Protein sequencing"
    ],
    [
      "plos",
      "/Social sciences/Psychology/Cognitive psychology/Language"
    ]
  ],
  "openalex": "W4416512467",
  "pdf_url": "https://journals.plos.org/ploscompbiol/article/file?id=10.1371/journal.pcbi.1014805&type=printable",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-28",
  "publisher": "Public Library of Science",
  "ris": "https://research.draconitic.com/paper/3pwtdp3z2h5a7.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:17:37Z",
      "native_id": "10.1371/journal.pcbi.1014805",
      "source": "plos",
      "url": "https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1014805"
    }
  ],
  "subjects": [
    {
      "field": "biochemistry",
      "origin": "plos:molecular biology",
      "subfield": "molecular-biology"
    },
    {
      "field": "biology",
      "origin": "plos:immunology",
      "subfield": "immunology"
    },
    {
      "field": "biology",
      "origin": "plos:neuroscience",
      "subfield": "neuroscience"
    }
  ],
  "title": "Attribution assignment for deep-generative sequence models enables interpretability analysis using positive-only data",
  "title_text": "Attribution assignment for deep-generative sequence models enables interpretability analysis using positive-only data",
  "type": "peer-reviewed",
  "updated_at": "2026-10-30T12:17:37Z",
  "url": "https://research.draconitic.com/paper/3pwtdp3z2h5a7",
  "venue": "PLOS Computational Biology",
  "versions": []
}