TAXyl

TAXyl predicts the optimum temperature of activity for xylanases belonging to glycoside hydrolase families 10 (GH10) and 11 (GH11), enabling in-silico thermal classification of these enzymes.

Key Features:

• GH family specificity: Targets xylanases from glycoside hydrolase families 10 (GH10) and 11 (GH11).
• Sequence-based Random Forest classifier: A sequence-based Random Forest model trained on a combination of various protein features to predict thermal optima.
• Thermal class distinction: Classifies xylanases into non-thermophilic, thermophilic, and hyper-thermophilic categories.
• Cross-validation evaluation: Model performance assessed using multiple iterations of six-fold cross-validation with a reported mean accuracy of approximately 0.79.
• Metagenomic screening capability: Applicable to targeted screening of metagenomic data to identify putative xylanases with specific thermal dependencies.
• Empirical discovery support: Has been used to identify three novel xylanases from sheep and cow rumen microbiota.

Scientific Applications:

• Novel enzyme discovery: Identification and prioritization of candidate xylanases from environmental and metagenomic datasets for experimental validation.
• Biocatalyst selection: Selection of xylanases with appropriate thermal optima for industrial processes such as lignocellulosic biomass degradation.
• Thermal characterization: In-silico classification to inform downstream biochemical characterization and enzyme engineering efforts.

Methodology:

Sequence-based Random Forest classifier trained on a combination of various protein features and evaluated using multiple iterations of six-fold cross-validation, yielding a mean accuracy of ~0.79.