Protein homology modelling from sequence to validated 3D model
No crystal structure for your target? ReactorMD finds suitable templates, builds a 3D model from your sequence and validates it, then reports honestly how far the model can be trusted.
Who this service is for
- Researchers working on proteins with no experimental structure
- Students preparing a receptor for docking or MD
- Groups studying a mutant or species variant
What to send us
- Target protein sequence (FASTA) and organism
- Optional preferred template, key residues or known ligands
- Intended use: docking, MD or structural analysis
How the work is done
- Search structure databases for related templates and compare coverage and identity.
- Align the target to the best template or templates.
- Build several candidate models.
- Refine loops and side chains.
- Validate with stereochemical checks, clash analysis and model quality scores, then select the best model.
What you receive
- Final 3D model in PDB format with alternative models if relevant
- Alignment and template details
- Validation figures and scores with a plain-language reliability statement
- Methods text for your report
Homology modelling packages
| Models Generated | Price (USD) | Estimated Turnaround |
|---|---|---|
| 1 validated model | $19.99 | Up to 12 hours |
| 3 validated models | $49.99 | Up to 24 hours |
| 5 validated models | $74.99 | Up to 24 hours |
| Additional model | $14.99 | Scope dependent |
What this method can and cannot tell you
Model quality depends mainly on template similarity. Below about 30 percent sequence identity, models are unreliable for detailed ligand work, and loops are always uncertain. Limitations are stated in the report.
Frequently asked questions
How accurate is a homology model?
Accuracy rises with sequence identity to the template. Above 50 percent identity the core is often reliable; below 30 percent be cautious about fine detail such as side-chain contacts.
Can the model be used for docking afterwards?
Yes, but binding-site residues should be checked. The model can then go straight into the docking or MD service.
What do you need to start?
The amino acid sequence in FASTA format and the organism. Tell us your intended use so the model is optimised for it.
Is a deep-learning predicted structure better?
Predicted structures can be excellent for many targets. The right choice depends on the protein, and the report explains which approach suits your case.
Related services
Online protein–ligand molecular docking: structure and ligand preparation, binding-site definition, pose ranking, interaction analysis and a full report.
ServiceMolecular dynamics simulationGROMACS molecular dynamics simulation of protein–ligand complexes with equilibration, production run and full trajectory analysis (RMSD, RMSF, Rg, SASA, H-bonds, PCA, FEL).
ServiceVirtual screeningStructure-based virtual screening of compound libraries: library preparation, high-throughput docking, filtering and a ranked, traceable hit list.
Learn more
Ready to start?
Tell us about your target and compounds. The price and scope are confirmed before payment.
Computational results are predictions and support, but do not replace, experimental validation. Page reviewed by ReactorMD, 16 October 2026.