SUBSEQ.BIO
DOCS-OVERVIEW

Subseq Overview

Subseq is a workbench for protein design and biomolecular computation. Upload structures or sequences, configure jobs in the UI, track progress, inspect results, and download the files you need for the next experiment.

It is built for scientists who want curated computational biology programs without rebuilding the same job plumbing for every target, dataset, and candidate campaign.

SubSeq RFdiffusion3 configuration with PDL1 structure preview

Capabilities

Structure predictionProtein, complex, ligand, nucleic-acid, and mixed biomolecular predictions.
Protein designBackbone generation, binder design, motif scaffolding, and sequence design.
Docking and scoringProtein-ligand docking, protein-protein docking, interface scoring, and metric filtering.
Simulation and figuresMolecular dynamics, minimization, and static result figures for review or communication.
SubSeq pipeline builder with RF Diffusion, LigandMPNN, Boltz-2, GROMACS, and PyMOL+ steps

Pipelines

Pipelines chain compatible jobs into a repeatable campaign. A typical design review can move from backbone generation to sequence design, structure prediction, relaxation, and a final figure without manually rebuilding each step.

Example: RF Diffusion generates candidate backbones, LigandMPNN designs sequences, Boltz-2 predicts and scores structures, GROMACS performs a relaxation or MD check, and PyMOL+ renders review images.

Read the pipeline guide.

  • Use pipelines after the individual steps and file handoffs are known to work.
  • Keep intermediate jobs available for inspection, filtering, and troubleshooting.
  • Use standalone jobs when a scientist should choose a specific file between steps.

Pricing

SubSeq uses pay-per-runtime credits. Each program has transparent pricing so you can estimate a run before submitting it, and completed jobs report the credits spent. The goal is to make exploratory one-off jobs and larger campaigns predictable without subscriptions or opaque usage tiers.

View current program pricing.

Guided Design Support
SubSeq also offers guided design support. Contact support@subseq.bio with inquiries.

Ways To Use SubSeq

The web app is the main interactive workspace: choose a program, upload or select data, set the canonical job configuration, then inspect status and outputs.

For automation, the same jobs and datasets are available through the REST API and MCP server. Teams let a lab share credits, jobs, datasets, projects, pipelines, and API keys.

  • Form fields match the structured parameters used by API and MCP submits.
  • Projects and memos keep exploratory runs findable later.
  • Advanced options stay available when a run needs lower-level control.
Boltz-2 result structure preview with selected protein residues

Program Guides

RFdiffusion3Design all-atom binders, de novo monomers, and symmetric oligomers with RFdiffusion3. Genie 3Generate de novo protein backbones, scaffold motifs, or design binders with Genie-compatible inputs. BoltzGenGenerate candidate protein or peptide binders for protein, peptide, or small-molecule targets. Protein HunterRun iterative protein design and triage workflows aimed at producing ranked candidate binders or designed proteins. FreeBindCraftRun de novo binder search workflows for target proteins, then inspect and triage generated candidate binders. RF DiffusionGenerate protein backbone structures from high-level structural constraints. LigandMPNNDesign protein sequences from backbone structures, with protein-only, ligand-aware, soluble, and membrane-aware model choices. ESMCExtract protein sequence embeddings and masked-token plausibility scores from a sequence-only protein language model. Evo 2Use a DNA language model for sequence generation, likelihood scoring, and embedding extraction. ESMFold2Predict all-atom biomolecular structures from sequence, FASTA files, or structured JSON inputs. Chai-1Predict biomolecular complex structures from a FASTA-like input describing proteins, ligands, DNA, RNA, glycans, and mixed systems. OpenFold3Predict structures for biomolecular assemblies containing proteins, RNA, DNA, and small molecules. OpenDDERun preview all-atom structure prediction for protein, nucleic-acid, ligand, ion, and mixed biomolecular systems. BioEmuSample monomer protein conformational ensembles from an A3M alignment. Boltz-2Predict protein, nucleic-acid, and small-molecule complex structures, with optional small-molecule affinity output. DiffDock-LPredict candidate small-molecule binding poses against a prepared protein structure. DiffDock-PPPredict rigid-body docking poses for one protein receptor and one moving protein partner. ESMFoldPredict protein structures quickly from amino-acid sequence without MSAs or templates. GROMACSRun molecular dynamics preparation, minimization, prepared-system execution, and checkpoint continuation workflows. OpenMMRun conventional energy minimization or short molecular dynamics from standard biomolecular structures or prepared systems. MMseqs2Search, cluster, and build multiple-sequence alignments from FASTA inputs. Shape-CompCalculate interface shape-complementarity style scores for protein complexes or candidate interfaces. FilterPost-process a completed batch of sequence or structure results by numeric metrics and keep the selected files together. Beautiful AtomsRender molecular and materials structures as static figures for inspection and communication. AlphaFold 2Predict protein monomer and multimer structures from FASTA inputs, with confidence outputs for local structure, domain placement, and interfaces. PyMOL+Render molecular still images from structure files using PyMOL-style selections and layers. UtilityRun small custom file-processing scripts against uploaded data, datasets, or previous outputs.