AI-Powered Enzyme Engineering Data for Biocatalyst Development

Pravin Kumar R published a dataset on figshare in 2026 detailing an AI-driven protein-engineering framework. The work describes a dataset of 1.39 million structural fragments used to predict productive enzyme substitutions for biocatalysis. It reports results from evaluating five AI-prioritized enzyme variants for the synthesis of sitagliptin.

Use Cases

• Training AI models for enzyme property prediction based on structural fragment data
• Benchmarking computational protein design methods against experimental validation results
• Analyzing the relationship between mutation combinations and catalytic efficiency/solubility
• Optimizing green chemistry parameters like solvent systems for industrial biocatalysis

Strengths

• Dataset includes a specific count of 1.39 million structural fragments
• Experimental validation includes concrete performance metrics: 51% conversion at 20% enzyme load, >99% enantiomeric excess, and 89% conversion upon scale-up
• Describes a specific, industrially relevant application: conversion of a prochiral ketone to sitagliptin

Limitations

• Dataset size is very small at 6.4 KB, indicating limited scope or summary-level data
• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment

Provenance

Source

figshare

Collection Method

Likely generated from computational protein design and experimental validation as described.

Freshness

Last updated 2026-05-14 12:34:42; freshness should be verified