Table 2 Performance of different technologies (array, WGS low coverage and WES) with different ROH size classes (Short < 1 Mb, Medium 1 – 8 Mb and Long > 8 Mb)

Short < 1 Mb

Poor performance due to low SNP coverage. Can be adjusted to detect ROH by modifying the number of SNPs required in a ROH.

Able to detect but need to build adjustment for genotype calling errors.

Able to detect but only in selected genomic regions. Software like H³M² allows meaningful regional analysis [39].

Detection of rare variants involved in deleterious recessive alleles and directional dominance [11, 12]. Analysis of LD patterns and extreme bottle necks [33].

Medium 1-8 Mb

Able to detect if the array has at least 300 K SNPs. ROH boundaries will be fuzzier in comparison with WGS low coverage data.

Good performance but need to build adjustment for genotype calling errors. Allowing 3 heterozygous SNPs per ROH would grant meaningful outcomes.

Able to detect, but only in selected genomic regions and boundaries of ROH could be fuzzy if they reach into non-exonic regions [49].

Detection of rare variants involved in diseases. Analysis of inbreeding depression. Genome architecture and ROH island detection [50]. Population history, bottle necks, remote consanguinity and genetic drift [51].

Long > 8 Mb

Good performance if the array has at least 300 K SNPs.

Good performance but need to build adjustment for genotype calling errors. Allowing 3 heterozygous SNPs per ROH would grant meaningful outcomes.

Poor performance due to short size of most exons and their sparsity across the genome.

Analysis of inbreeding depression. Validation of GWAS findings. Population history and cultural practices, close consanguinity [6, 41].