iT3SE-PX

iT3SE-PX predicts bacterial type III secreted effectors (T3SEs) from protein sequences to support studies of host–pathogen interactions.

Key Features:

• Sequence-Based Identification: Predicts T3SEs solely from protein sequence information.
• Feature Extraction from PSSM Profiles: Extracts three distinct types of features from position-specific scoring matrix (PSSM) profiles.
• XGBoost Feature Selection: Uses extreme gradient boosting (XGBoost) to rank features by classification importance and select optimal subsets.
• SVM Classification: Trains a support vector machine (SVM) classifier on the selected features to classify proteins as T3SEs or non-T3SEs.
• Performance Validation: Validated on two benchmark datasets using 100-time randomized 5-fold cross-validation and independent testing.

Scientific Applications:

• Host–pathogen interaction analysis: Supports analysis of bacterial host–pathogen interactions by identifying candidate T3SEs.
• Discovery of novel effectors: Predicts putative novel type III secreted effectors for experimental follow-up.
• Therapeutic target prioritization: Aids prioritization of candidate therapeutic targets among bacterial effectors.

Methodology:

Feature extraction from PSSM profiles (three feature types), feature ranking and selection using XGBoost, classification with SVM, and performance evaluation via 100-time randomized 5-fold cross-validation and independent testing on two benchmark datasets.