Improved genomic island predictions with IslandPath-DIMOB.

Area of Science:

• Genomics
• Bioinformatics
• Microbial Evolution

Background:

• Genomic islands (GIs) are crucial for bacterial and archaeal genome evolution and adaptability.
• GIs are enriched in virulence factors, antimicrobial resistance genes, and metabolic pathways, making them medically and industrially significant.
• Accurate prediction tools are needed for large-scale microbial genome analysis.

Purpose of the Study:

• To improve the accuracy and performance of the IslandPath-DIMOB genomic island prediction tool.
• To enhance the prediction of genomic islands, which play a key role in microbial adaptability.

Main Methods:

• Modified the decision algorithm for sequence composition biases.
• Updated the database of HMM profiles for associated mobility genes.
• Assessed accuracy using reference and manually curated GI datasets, comparing IslandPath-DIMOB v1.0.0 with other software.

Main Results:

• IslandPath-DIMOB v1.0.0 shows an 11.7% increase in recall and a 5.3% increase in precision compared to v0.2.0.
• Achieved the highest Matthews correlation coefficient among individual prediction methods tested.
• Demonstrated high recall (46.9%) at high precision (87.4%).

Conclusions:

• The enhanced IslandPath-DIMOB v1.0.0 facilitates more accurate studies of genomic islands.
• This tool supports research into the roles of GIs in microbial adaptability across medical, environmental, and industrial fields.
• The improved tool is freely available via the IslandViewer webserver and as standalone software.