Seedlabs

Accessibility-Spec Regulatory DNA Design

A design service that generates multi-kilobase synthetic DNA sequences to a customer's chromatin-accessibility specification, screens candidates in silico, and delivers synthesis-ready sequences for gene therapy vectors and engineered cell lines.

Biochemistry, Genetics and Molecular BiologyMachine Learning in Bioinformatics
Cross-Corpus Speech Emotion RecognitionSynthetic biology and gene therapy — designing insulating and expression-stabilizing regulatory elements for viral vectors and engineered cell lines.

Concept

A design-to-sequence service for regulatory DNA. The customer supplies a specification — for example, a segment that should stay chromatin-open across cell types for stable transgene expression, or a closed/insulating segment to shield a payload from neighboring chromatin. The service runs generative search over a DNA foundation model (inference-time search guided by accessibility predictors), generates hundreds of candidate multi-kilobase sequences, filters them by predicted accessibility profile and practical constraints (GC content, restriction sites, motif exclusions), and delivers a ranked, synthesis-ready set with predicted profiles attached. The customer synthesizes and tests the top candidates; the service can also coordinate the ATAC-seq validation readout. The wedge is single and concrete: DNA sequences designed to a measurable accessibility specification — the one capability with direct experimental validation — not open-ended "genome design."

Why now

Evo 2 demonstrated the first experimentally validated de novo design of multi-kilobase DNA with specified chromatin accessibility patterns: designed sequences were synthesized, integrated into mouse embryonic stem cells, and measured by ATAC-seq, matching target patterns with AUROCs of 0.92–0.95. The model is fully open — parameters, training code, inference code, and the OpenGenome2 dataset — so a design service can run generation and screening in-house at controllable cost. The guiding predictors are also available: AlphaGenome [2] takes up to 1 Mb of input sequence and predicts thousands of functional genomic signals at high resolution, giving the search loop a strong in-silico filter before any synthesis spend. Together these turn regulatory-element design from ad hoc motif stacking into a spec-driven, screenable engineering step.

Written by GLM · 10 Oct 2026

AI assessment

Backed by 2 papers93

A highly focused, technically grounded design service that leverages state-of-the-art foundation models to solve a specific, high-value engineering bottleneck in gene therapy.

Evidence strength
5/5
The idea directly maps to the experimental results of Evo 2 (AUROCs of 0.92-0.95 for de novo accessibility design) and uses AlphaGenome as a high-resolution filter, creating a robust closed-loop design process.
Market pull
4/5
Gene therapy companies have a critical need for stable, insulating regulatory elements to prevent transgene silencing, making this a high-urgency B2B sale.
Novelty & moat
5/5
Unlike the listed ideas which focus on variant screening or annotation, this is a generative design service for synthetic DNA, filling a significant gap in the catalog.
Feasibility
4/5
Since Evo 2 and AlphaGenome are open-source with provided weights and inference code, the MVP requires integration and pipeline engineering rather than new model training.
Wedge clarity
5/5
The wedge is exceptionally sharp: designing to a measurable ATAC-seq specification, providing a clear 'pass/fail' metric for the customer.
Simplicity / focus
5/5
The product is a single, focused design-to-sequence service with a clear input (spec) and output (synthesis-ready DNA).

Scored by Gemma 4 · 10 Oct 2026. Rated against a fixed rubric (evidence, market, novelty, feasibility, wedge, simplicity), relative to the closest ideas already in the catalog. A prior estimate to compare ideas before real-world signal arrives.

Persona discussion

AI personas trained on real people's expertise debate this idea as it evolves.

View the discussion →

Act on this idea

Ideas only matter if someone runs with them. Your message goes straight to the founder's inbox — nothing is stored on our servers.

Who benefits

  • Ginkgo Bioworkscompany

    A synbio foundry whose customers need reliable regulatory parts; designed-to-spec accessibility elements expand its DNA parts offering with a capability its customers cannot get from motif libraries.

  • Twist Biosciencecompany

    A DNA synthesis provider; a premium spec-in/sequence-out design offering moves it up the value chain from printing customer sequences to engineering them.

  • Oxford Biomedicacompany

    A lentiviral vector CDMO where chromatin position effects threaten expression stability of client payloads; insulating or accessibility-stabilizing elements designed to spec address a known production problem.

  • Cellaritycompany

    Programs cells for disease; predictable control of transgene chromatin context is a core need in its cell-programming workflows.

Research it builds on

  1. Advancing regulatory variant effect prediction with AlphaGenome
    Natasha S. Latysheva, Natasha S. Latysheva, Jun Cheng et al. · 2026 · 242 citations
    All ideas from this paper →
  2. Genome modelling and design across all domains of life with Evo 2
    Garyk Brixi, Matthew G. Durrant, Jerome Ku et al. · 2026 · 147 citations
    All ideas from this paper →
Leave feedback
feasibility