3.2 High genetic affinity between modern goat populations in Southwest Asia and East-Africa
To determine genome-wide introgression between present-day domestic groups, D and F -branch (f b) statistics were calculated with Dsuite software (V3, Malinsky et al. 2021). The results show strong signals of introgression between IRQ-IRN (f b = 0.61, Z=35.21), EAFR-IRQ (f b = 0.419, Z = 14.32) and EAFR -IRN (f b = 0.374, Z =11.98) populations (Figure 2a ). Additional cross-drainage introgression was also detected between some trios such as between common ancestor of African and both IRQ (f b=0.58, Z=12.32) and IRN (f b=0.51, Z=9.58) populations, while we did not find evidence of gene flow between WAFR goats and SWA populations. These values are higher than the EAFR-IRQ and EAFR-IRN values and may suggest that the latter values reflect introgression of the common ancestor of EAFR and WAFR in the Levant region. The f b results also highlighted excessive allele sharing between CHN/BNG, PKS/IRN and EUR/WAFR populations (Figure 2a ). The results were further supported by TreeMix analysis which reveals signals of genomic introgressions from EAFR goats into IRQ (m = 1) and IRN (m = 2), explaining 99.7% and 99.9% variance, respectively (Figure 2b and Figure S7 ).
Quantifying gene flow of EAFR goats into both IRQ and IRN populations by using the Loter program (Dias-Alves et al. 2018) shows that the proportion of introgression was not uniform among present-day goats in different parts of SWA region. The frequency of gene flow gradually decreased with the increase of distance from EAFR (Figures 2c and 2d ). The average frequency was highest (~65%) and lowest (~36%) in samples located in IRQ and south parts of IRN, respectively. Using the f d statistic (Malinsky et al. 2021) we further localize genomic regions with significant probability of introgression from EAFR goats into both IRQ and IRN populations (Figures S8a and S8b ). The length of the introgressed regions varies from several hundreds to more than 200,000 base pairs. Selecting the top 1% of these regions, we obtained 2456 (2218 genes) and 974 (966 genes) autosomal regions as high-confidence introgressed segments for IRQ and IRN goat groups, respectively (Tables S4 and S5 ).