AI-Generated PHA Synthase Sequences with In Vivo Validation

Approximately 10,000 novel Polyhydroxyalkanoate synthase protein sequences designed using a conditional variational autoencoder trained on native sequences from UniProt. 16 sequences were selected for in vivo validation based on conserved residues and structural features, with 2 confirmed active in yeast S. cerevisiae. The dataset was created by Tuula Tenkanen and last updated in March 2026.

Use Cases

• Evaluating AI-generated enzyme designs based on sequence conservation criteria
• Training models for protein sequence generation based on native enzyme data
• Analyzing structural features like helix count for enzyme activity prediction
• Benchmarking in vivo validation results for novel protein sequences

Strengths

• Includes 16 sequences selected for in vivo validation with specific selection criteria
• Contains 2 novel enzymes confirmed active with 87 and 98 amino acid substitutions
• Model trained on native protein sequences from UniProt

Limitations

• Row count is unknown, which may limit suitability assessment
• Column-level documentation is absent; field semantics must be inferred after download
• Dataset size is 5.5 KB, indicating a very limited scope

Provenance

Source

figshare

Collection Method

Sequences generated by a conditional variational autoencoder trained on UniProt data.

Freshness

Last updated 2026-03-19 17:34:36