GIPSy2: high-performance and scalable genomic island prediction software

Abstract

Abstract Dealing with genomic mobility is a complex task for current predictors. With an increasing number of sequencing genomes, there is a constant demand for software that can handle multiple inputs. Considering this, we present the Genomic Island Prediction Software 2 (GIPSy2), a new version of well-established software for predicting bacterial genomic islands and mobilome. Statistical methods were used to provide the values associated with each prediction, such as Fisher’s exact test, Support vector machine, and Logistic regression. The new version also improves scalability, allowing the simultaneous analysis of multiple genomes, and provides structured outputs to facilitate interpretation and reproducibility. Comparative analyses show that GIPSy2 achieves performance comparable to the original version under default settings, while offering increased flexibility through user-defined parameterization. These improvements make GIPSy2 a versatile tool for genomic island prediction across diverse bacterial datasets. GIPSy2 is currently available on Zenodo repository at https://zenodo.org/doi/10.5281/zenodo.10222587 .