SLPFA

SLPFA predicts the subcellular localization of proteins from amino acid sequences to aid functional annotation of proteins and genes.

Key Features:

• Sequence alignment and composition integration: Combines sequence alignment information with feature vectors derived from amino acid composition (frequency).
• Support Vector Machines: Implements classification using support vector machines (SVMs).
• Evaluation dataset: Demonstrated on plant datasets extracted from the TargetP database.
• Cross-validation performance: Reported overall accuracy of 0.9096 and average Matthews correlation coefficient (MCC) of 0.8655 using fivefold cross-validation.
• Sequence-only prediction: Operates without relying on gene ontology data, enabling predictions for proteins lacking auxiliary annotations.

Scientific Applications:

• Subcellular localization prediction: Predicts the cellular compartments of proteins directly from amino acid sequence data.
• Functional annotation: Supports functional annotation of proteins and genes by providing localization information.
• Plant proteome analysis: Applied to plant protein datasets for localization studies using sequences from TargetP.
• Annotation of novel proteins: Enables localization prediction for newly discovered proteins that lack Gene Ontology annotations.

Methodology:

Integrates sequence alignment techniques with amino acid-frequency-derived feature vectors and classifies proteins using support vector machines, with performance assessed by fivefold cross-validation on plant datasets from the TargetP database.