SURE-Pipe Benchmarking datasets

Simulations were generated across genome sizes ranging from 100 kb to 75 Mb. For accuracy assessment, 150 independent simulations using 4 Mb genomes were performed, and the outputs of SURE-Pipe v1.1 were compared with KEC v1.1 and FUR v4.3 using nucleotide-level sensitivity, specificity, precision, accuracy, and Matthews correlation coefficient (MCC). Computational scalability was further assessed through genome-size and genome-number scaling experiments.

Genome-size benchmarking used datasets of five target and five neighbouring genomes ranging from 0.1 Mb to 640 Mb. Genome-number benchmarking fixed genome size at 5 Mb while increasing target:neighbour datasets from 5T:20N up to 160T:640N genomes. Runtime and peak memory usage were recorded using /usr/bin/time on a Pop!_OS 22.04 LTS workstation equipped with an Intel® Xeon® E-2124G CPU (4 cores, 3.40 GHz) and 16 GB RAM.

The pairwise genome comparison mode was validated using six genome pairs spanning diverse taxonomic groups, genome sizes (10.7 kb–69.5 Mb), and GC contents (27–76%), including viral, bacterial, and fungal genomes. Additionally, groupwise genome comparison was performed across 24 closely related Bacillus species to identify species-specific genomic regions. Four genomes per species (>97% ANI) were selected, while neighbouring datasets included reference genomes from 42 Bacillus species.

This analysis enabled the identification of conserved intraspecies regions, species-specific unique regions, and regions shared among neighbouring taxa.