BACKGROUND : Genomic islands (GIs) are inserts of foreign DNA that have potentially arisen through horizontal gene
transfer (HGT). There are evidences that GIs can contribute significantly to the evolution of prokaryotes. The
acquisition of GIs through HGT in eukaryotes has, however, been largely unexplored. In this study, the previously
developed GI prediction tool, SeqWord Gene Island Sniffer (SWGIS), is modified to predict GIs in eukaryotic
chromosomes. Artificial simulations are used to estimate ratios of predicting false positive and false negative GIs by
inserting GIs into different test chromosomes and performing the SWGIS v2.0 algorithm. Using SWGIS v2.0, GIs are
then identified in 36 fungal, 22 protozoan and 8 invertebrate genomes.
RESULTS : SWGIS v2.0 predicts GIs in large eukaryotic chromosomes based on the atypical nucleotide composition of
these regions. Averages for predicting false negative and false positive GIs were 20.1% and 11.01% respectively. A
total of 10,550 GIs were identified in 66 eukaryotic species with 5299 of these GIs coding for at least one functional
protein. The EuGI web-resource, freely accessible at http://eugi.bi.up.ac.za, was developed that allows browsing the
database created from identified GIs and genes within GIs through an interactive and visual interface.
CONCLUSIONS : SWGIS v2.0 along with the EuGI database, which houses GIs identified in 66 different eukaryotic
species, and the EuGI web-resource, provide the first comprehensive resource for studying HGT in eukaryotes.
Additional file 1: Table S1. Eukaryotic organisms used for database
construction by GI prediction with SWGIS v2.0. The table presents all fungal,
protozoan and invertebrate species that were used for GI prediction in this
study along with the amount of GIs predicted in each species and the
amount of genes retained within these GIs.