ShadowCaster

ShadowCaster detects horizontal gene transfer (HGT) events in prokaryotic genomes by applying support vector machines to nucleotide composition and refining predictions with phylogenetic models without tree reconstruction to improve detection of both close and distant transfers.

Key Features:

• Hybrid Approach: Integrates nucleotide composition-based predictions with phylogenetic-model-based refinement in a sequential manner using support vector machines (SVMs) for initial predictions.
• Enhanced Accuracy: Improves detection accuracy for both close and distant HGT events and has demonstrated performance identifying HGT related to heavy metal resistance in Rhodanobacter denitrificans.
• Scalability and Efficiency: Handles large genomic datasets more efficiently than computationally intensive phylogenetic reconstruction methods.

Scientific Applications:

• Evolutionary analysis of HGT: Detects gene transfer events to study prokaryotic genome evolution and the spread of genetic material across lineages.
• Microbial ecology and gene flow: Maps gene flow within microbial communities to investigate ecological interactions and horizontal dissemination of functions.
• Adaptive trait identification: Identifies transferred genes associated with adaptive phenotypes, such as heavy metal resistance exemplified in Rhodanobacter denitrificans.

Methodology:

Performs initial HGT prediction using support vector machines on nucleotide composition, then refines predictions with phylogenetic models applied without tree reconstruction in a sequential hybrid workflow.