DCAlign v1.0: aligning biological sequences using co-evolution models and informed priors.

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Multiple sequence alignments (MSAs) are crucial for understanding protein evolution and function.
• Identifying conserved and co-evolving residues is key to accurate MSAs.
• Existing alignment methods face computational challenges with pre-processing steps.

Purpose of the Study:

• To introduce DCAlign v1.0, a novel alignment method.
• To address the computational demands of pre-processing in sequence alignment.
• To improve the efficiency of aligning homologous sequences.

Main Methods:

• DCAlign utilizes conservation and co-evolution signals in homologous sequences.
• The method incorporates an empirical prior over variables representing insertions and deletions.
• Implementation in Julia facilitates accessibility and further development.

Main Results:

• DCAlign v1.0 demonstrates a significant reduction in overall computing time.
• The empirical prior effectively models insertions and deletions, enhancing alignment efficiency.
• The method maintains the ability to leverage conservation and co-evolutionary information.

Conclusions:

• DCAlign v1.0 offers a computationally efficient approach to multiple sequence alignment.
• The integration of an empirical prior is key to reducing processing time.
• DCAlign provides a valuable tool for genomic and evolutionary analyses.