ProteoKnight: Convolution-based phage virion protein classification and uncertainty analysis

TL;DR

ProteoKnight introduces a novel image-based encoding for phage virion protein classification using CNNs, achieving high accuracy and enabling uncertainty estimation through Monte Carlo Dropout, thus improving annotation reliability.

Contribution

The paper presents a new encoding method for protein sequences that enhances spatial information retention and applies pre-trained CNNs for accurate PVP classification with uncertainty analysis.

Findings

Achieved 90.8% accuracy in binary PVP classification.

Encoding method outperforms FCGR by reducing spatial information loss.

Uncertainty analysis reveals confidence variability based on protein class and length.

Abstract

\textbf{Introduction:} Accurate prediction of Phage Virion Proteins (PVP) is essential for genomic studies due to their crucial role as structural elements in bacteriophages. Computational tools, particularly machine learning, have emerged for annotating phage protein sequences from high-throughput sequencing. However, effective annotation requires specialized sequence encodings. Our paper introduces ProteoKnight, a new image-based encoding method that addresses spatial constraints in existing techniques, yielding competitive performance in PVP classification using pre-trained convolutional neural networks. Additionally, our study evaluates prediction uncertainty in binary PVP classification through Monte Carlo Dropout (MCD). \textbf{Methods:} ProteoKnight adapts the classical DNA-Walk algorithm for protein sequences, incorporating pixel colors and adjusting walk distances to capture intricate protein features. Encoded sequences were classified using multiple pre-trained CNNs. Variance and entropy measures assessed prediction uncertainty across proteins of various classes and lengths. \textbf{Results:} Our experiments achieved 90.8% accuracy in binary classification, comparable to state-of-the-art methods. Multi-class classification accuracy remains suboptimal. Our uncertainty analysis unveils variability in prediction confidence influenced by protein class and sequence length. \textbf{Conclusions:} Our study surpasses frequency chaos game representation (FCGR) by introducing novel image encoding that mitigates spatial information loss limitations. Our classification technique yields accurate and robust PVP predictions while identifying low-confidence predictions.