AI-Predicted Enzyme Variants for Sitagliptin Synthesis

1.39 million structural fragments were used to train an AI-driven protein-engineering framework. The dataset, authored by Pravin Kumar R and last updated in May 2026, contains AI-prioritized enzyme variants engineered for the biocatalytic synthesis of sitagliptin. Five variants, each containing nine mutations, were validated at a 7 L fermentation scale.

Use Cases

• Training machine learning models for protein function prediction based on interaction energy descriptors and solvent effects.
• Benchmarking AI-driven enzyme engineering frameworks against experimental screening results.
• Analyzing the relationship between mutation combinations and catalytic stability in transaminases.
• Evaluating the performance of engineered enzymes under different solvent systems, such as ethanol/PEG-400.
• Optimizing industrial biocatalysis processes for enantioselective synthesis based on conversion rates and enzyme load data.

Strengths

• Dataset is derived from a framework integrating 1.39 million structural fragments.
• Five AI-prioritized variants were experimentally validated at a 7 L fermentation scale.
• Results include specific performance metrics, such as 89% conversion upon scale-up and >99% enantiomeric excess.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is very small at 7.4 KB, indicating limited scope.

Provenance

Source

Pravin Kumar R

Collection Method

Likely contains results from an AI-driven 6D-grid protein-engineering framework integrating interaction energy descriptors and solvent effects.

Freshness

Last updated 2026-05-14 12:34:43; freshness should be verified.