Enzyme engineering
Design enzyme variants worth making
Map conserved positions, predict kinetics, and score mutations so only the most promising variants reach the bench.
Trusted by 16,500+ researchers and published in
Place the enzyme in its family
Search for homologs and align the family to see which positions are conserved and which co-evolve. Those are the residues worth keeping, and the ones worth mutating.
Design variants and scaffolds
Generate new scaffolds or redesign sequences around the active site, then rank variants by predicted or measured fitness before committing to a library.
Check stability and validate
Compare thermostability, solubility and structural geometry across your variants, so only the ones likely to fold and express reach the assay.
Enzyme engineering use cases
Pick the method that fits your question and evidence. Each one comes with a workflow you can open and edit.
Enzyme design
Build protein scaffolds around catalytic geometry, redesign ligand-aware sequences, and connect computation to biochemical testing.
Sequence alignment
Align two sequences, a gene family or a protein family online with Clustal Omega, MAFFT and MUSCLE5, then compare the alignments to see which columns you can trust.
Inverse folding
Design amino-acid sequences for a fixed protein backbone, compare complementary models, and refold candidates before selection.
Molecular docking
Dock a ligand with AutoDock Vina, GNINA or DiffDock-L, or two proteins with LightDock. Every pose keeps its own score, and PoseBusters checks the geometry before you rely on it.
Guides for enzyme engineering
Background guides on the methods, scores, and numbers behind these workflows.

Proteins · 4 min
What is the fastest enzyme?
Catalase is usually named the fastest enzyme: one catalase molecule is estimated to decompose 16 million to 44 million hydrogen peroxide molecules per second. Carbonic anhydrase has the fastest well-measured active sites, at 1.4 million to 4.4 million CO2 molecules per second.

Structures · 17 min
What is pLDDT, and what counts as a good score
pLDDT is the per-residue confidence score that AlphaFold, ESMFold and Boltz attach to predicted structures. Learn what the 90, 70 and 50 cutoffs mean and why neither a high nor a low score settles whether a region is real.

Structures · 19 min
TM-score vs RMSD
RMSD and TM-score both compare two protein structures, but they answer different questions. Learn how each is calculated, what values count as good, and which one to use for prediction, fold search, docking or design.

Proteins · 18 min
What is a protein language model, and what is it used for
A protein language model is a neural network trained to predict amino acids from the rest of a sequence. Learn how ESM-2, ESM C, ESM3, ProtT5 and ProGen2 differ, what their embeddings and likelihoods are used for, how they compare with alignment-based methods like AlphaFold2, and which licences allow commercial use.
FAQs
ProteinIQ supports the computational stages of enzyme engineering in one connected workspace. Researchers can annotate sequences, inspect substrate context, predict turnover or stability, model structures, and score variants while keeping every output tied to the enzyme candidate it describes. Begin with an enzyme sequence or structure and select a workflow for activity, stability, mutation, or substrate analysis. ProteinIQ returns the source tables, structures, scores, and files needed to compare candidates before experimental enzyme assays.
ProteinIQ supports enzyme engineering workflows for enzyme annotation, substrate context review, activity prediction, kcat-style scoring, mutation and stability screening, structure modeling, and inhibitor or substrate triage. The platform keeps upstream tool outputs available so enzyme decisions can be reviewed rather than treated as black-box results.
ProteinIQ enzyme tools commonly accept enzyme FASTA sequences, PDB structures, substrate SMILES, inhibitor libraries, mutation lists, and tabular inputs depending on the upstream method. Each workflow keeps the accepted input type tied to the tool that generates the activity, stability, or structure output.
Yes. ProteinIQ can compare enzyme variants with mutation, stability, solubility, structure, and activity-related outputs depending on the selected tools. The results are preserved as tables, structure files, logs, and other artifacts so candidate variants can be compared side by side.
ProteinIQ can run enzyme activity and kcat-oriented prediction tools when the required sequence, structure, or substrate inputs are available. These outputs are computational prioritization signals, so ProteinIQ presents the scores and input context for review rather than treating them as measured kinetic constants.
Yes. ProteinIQ can connect enzyme sequences or structures to design, mutation scoring, and stability prediction tools that help prioritize variants for testing. The platform keeps the redesigned sequences, mutation tables, stability scores, and structure evidence tied to the original enzyme context.
Yes. When an enzyme workflow includes compound context, ProteinIQ can help review substrate or inhibitor candidates with structure, docking, property, and ADMET-style evidence depending on the tools selected. These results support triage and experimental planning rather than replacing biochemical validation.
No. ProteinIQ provides computational enzyme engineering and prioritization evidence. Enzyme kinetics, activity, specificity, stability, expression, and substrate-scope claims still require experimental enzyme assays before they should be treated as validated results.
Yes. ProteinIQ enzyme workflows can export CSV score tables, sequence files, predicted or prepared structures, mutation outputs, docking or compound tables, logs, and upstream result artifacts depending on the tool. Exported files preserve the evidence needed for review outside ProteinIQ.
Start in ProteinIQ with an enzyme sequence or structure and choose either a workflow template or a specific tool for activity prediction, stability review, structure modeling, or compound triage. Running a small test input first is the best way to confirm formatting before scaling to larger variant sets.
AI answers from our docs. No sign-in needed.
Focus on research, not infrastructure
Start free with 200 credits a month and no installs. Upgrade when your project needs more compute.