{
  "abstract": "Proteins are fundamental to biological processes, with their function determined by the complex interplay between the amino acid sequence and the three-dimensional structure. Developing generative models capable of understanding this intrinsically multi-modal relationship is crucial for fields like drug discovery and protein engineering. Existing models often rely on a multi-stage training process where autoencoders that tokenize data into latent representations are trained in a first stage. Secondly, a generative model is trained on the latent representation of the autoencoder(s), i.e., generative modeling in a latent space. We hypothesize that this multi-stage training is not necessary to obtain performant co-design models and thus present SimpleDesign, an effective multi-modal protein design model trained directly in the data space. SimpleDesign leverages a single-stage end-to-end objective that combines discrete cross-entropy for sequences and a regression objective for structures. In order to effectively model the difference in sequence and structure modalities, we develop a Mixture-of-Transformer architecture that allows modality-specific processing while keeping global self-attention over both modalities. We train SimpleDesign on over 2M sequence-structure pairs achieving strong performance across co-design and unconditional sequence/structure generation benchmarks.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": "Journal reference: Transactions on Machine Learning Research, 08/2026",
  "arxiv": "2609.03377",
  "author_total": 7,
  "authors": [
    {
      "name": "Jiarui Lu"
    },
    {
      "name": "Yuyang Wang"
    },
    {
      "name": "Yizhe Zhang"
    },
    {
      "name": "Jiatao Gu"
    },
    {
      "name": "Navdeep Jaitly"
    },
    {
      "name": "Joshua M. Susskind"
    },
    {
      "name": "Miguel Ángel Bautista"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/12a9eeccqsah9.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2609.03377",
  "fulltext_url": null,
  "id": "12a9eeccqsah9",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2609.03377",
  "language": null,
  "licence": "LicenseRef-arXiv-nonexclusive-1.0",
  "licence_label": "arXiv licence",
  "licence_source": "article",
  "licence_url": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "q-bio.BM"
    ],
    [
      "openalex",
      "1312|Protein Structure and Dynamics"
    ]
  ],
  "openalex": "W7208765091",
  "pdf_url": "https://arxiv.org/pdf/2609.03377",
  "pmcid": null,
  "pmid": null,
  "published": "2026-09-03",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/12a9eeccqsah9.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:18:37Z",
      "native_id": "2609.03377",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2609.03377"
    }
  ],
  "subjects": [
    {
      "field": "biochemistry",
      "origin": "arxiv:q-bio.BM",
      "subfield": "biophysics"
    }
  ],
  "title": "SimpleDesign: A Joint Model for Protein Sequence and Structure Codesign",
  "title_text": "SimpleDesign: A Joint Model for Protein Sequence and Structure Codesign",
  "type": "preprint",
  "updated_at": "2026-10-30T12:18:37Z",
  "url": "https://research.draconitic.com/paper/12a9eeccqsah9",
  "venue": "arXiv",
  "versions": []
}