AlexSys: a knowledge-based expert system for multiple sequence alignment construction and analysis.

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Multiple sequence alignment (MSA) is crucial for analyzing conservation and diversity in biological systems.
• Existing MSA algorithms and meta-methods have limitations in accuracy and efficiency.
• Meta-methods often require running multiple aligners and selecting the best solution post-hoc, leading to inefficiency.

Purpose of the Study:

• To develop a novel expert system, AlexSys, for automated and efficient multiple sequence alignment of protein sequences.
• To create an intelligent inference engine within AlexSys for a priori selection of optimal alignment algorithms.
• To improve the balance between alignment quality and computational time for high-throughput biological data analysis.

Main Methods:

• Development of AlexSys, an expert system featuring an intelligent inference engine for automated aligner selection.
• Training the inference engine using a machine learning approach on a large dataset of reference multiple alignments.
• Evaluation of AlexSys performance on a test set of 178 protein sequence alignments.

Main Results:

• AlexSys successfully selects appropriate aligners a priori based on input sequence characteristics.
• The expert system demonstrates a favorable trade-off between the quality of the generated multiple sequence alignments and the computational resources required.
• AlexSys proves suitable for high-throughput bioinformatics projects requiring efficient and accurate sequence alignment.

Conclusions:

• AlexSys provides an efficient and effective solution for multiple sequence alignment of protein data.
• The intelligent, a priori aligner selection mechanism in AlexSys overcomes the inefficiencies of previous meta-methods.
• AlexSys is a valuable tool for researchers engaged in large-scale molecular biology and bioinformatics studies.