Seedlabs

Twilight-Zone Tree Service: structure-based phylogeny for divergent viruses

An analysis service that uses protein shapes to map how divergent viruses are related.

A hosted analysis service that turns a set of protein structures (e.g., viral glycoproteins from hundreds of isolates) into a confidence-scored multiple structure alignment and a phylogenetic tree, resolving evolutionary relationships that sequence comparison cannot detect.

Biochemistry, Genetics and Molecular BiologyProtein Structure and Dynamics
Microbial Symbiosis and ResistancePublic-health virology: arbovirus surveillance, isolate classification, and vaccine antigen selection

Concept

A turnkey service for virology and evolutionary labs: upload 50–5,000 predicted or experimental structures of one protein family — for example envelope glycoproteins from flavivirus isolates — and receive a reproducible analysis package: a multiple structure alignment with per-residue confidence scores, a structure-derived phylogenetic tree, and an interactive visualization for exploring clades. The alignment engine is open-source (FoldMason), but the product is the derived workflow the free webserver does not provide: batch runs over isolate collections, metadata curation, tree inference, and exportable, auditable reports for publication or surveillance use. The single wedge: evolutionary classification of proteins whose sequences have diverged past the "twilight zone" (~20–25% identity), where sequence-based trees lose signal.

Why now

Paper [0] demonstrates exactly this use case: using Flaviviridae glycoproteins, it shows that FoldMason's multiple structural alignments support phylogenetic analysis below the twilight zone. Critically, it makes the workflow practical at scale — aligning hundreds of thousands of structures, two orders of magnitude faster than other MSTA methods, with alignment quality exceeding state-of-the-art and built-in confidence scores — so an entire isolate collection or protein family can be analyzed in one run instead of being subsampled. Structure is conserved beyond sequence, and with predicted structures now abundant, the historical barrier was MSTA speed and quality; paper [0] removes both.

Written by GLM · 12 Oct 2026

AI assessment

Backed by 1 paper92

A highly focused, technically grounded service that solves a specific 'twilight zone' bottleneck in virology by productizing a high-performance structural alignment tool.

Evidence strength
5/5
The idea directly implements the specific use case (Flaviviridae glycoproteins) and performance claims (speed and accuracy) demonstrated in the provided research paper.
Market pull
4/5
Public health agencies and vaccine developers have a high urgency for accurate isolate classification, especially for divergent viruses where sequence-based methods fail.
Novelty & moat
4/5
While similar to genomic tree services, this shifts the mechanism from sequence to structure, filling a distinct gap for divergent proteins that existing genomic tools cannot address.
Feasibility
5/5
The core engine (FoldMason) is already open-source and proven; the MVP requires building a wrapper for batch processing, metadata management, and reporting.
Wedge clarity
5/5
The 'twilight zone' (20-25% identity) is a sharp, scientifically defined entry point that clearly differentiates the product from standard sequence aligners.
Simplicity / focus
5/5
The product is a single, cohesive workflow (upload structures -> get tree/alignment) without unnecessary feature creep or platform bloat.

Scored by Gemma 4 · 12 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

  • MRC-University of Glasgow Centre for Virus Researchorganization

    A leading flavivirus evolution research center that would use structure-based trees to publish relationships among divergent glycoproteins that sequence trees cannot resolve.

  • NIH Vaccine Research Centerorganization

    Selects vaccine antigens using structure-based reasoning; structure-derived phylogenies of viral surface proteins directly inform antigen and clade-coverage decisions.

  • CDC Division of Vector-Borne Diseasesorganization

    Tracks dengue, Zika, and West Nile viruses; a structure-based phylogeny service helps classify divergent field isolates that defeat sequence-based typing.

  • World Health Organizationorganization

    Global arbovirus surveillance needs to classify divergent flavivirus isolates; structure-based trees resolve relationships where sequence typing fails, supporting outbreak source attribution.

  • Takedacompany

    Markets the dengue vaccine Qdenga; glycoprotein structural phylogenies across flavivirus serotypes inform tetravalent antigen selection and coverage monitoring.

Research it builds on

  1. Multiple protein structure alignment at scale with FoldMason
    Cameron L. M. Gilchrist, Milot Mirdita, Martin Steinegger · 2026 · 114 citations
    All ideas from this paper →
Leave feedback
feasibility