Analysis of relationships by COLONY 2.0 and Sequoia
We conducted kinship analysis using COLONY 2.0 and Sequoia and found that Sequoia produced lower reliability of kinships for all types of relationships compared to COLONY 2.0. This is because COLONY 2.0 estimates kinships by calculating the likelihood for the entire set of individuals (Wang, 2004, 2012), whereas Sequoia identifies the most likely kinship based on a threshold likelihood ratio between the first and second most likely relationships for a given pair (Huisman, 2017), which is less powerful than the COLONY 2.0 calculation. Moreover, the target population in this study was considered to be highly inbred, making COLONY 2.0 more suitable for this study as it can account for inbreeding. Nevertheless, almost all the pairs estimated as FS or HS by Sequoia were similarly estimated by COLONY 2.0, whereas four out of eight UANN pairs were estimated only by Sequoia. This can be attributed to COLONY 2.0 originally lacking a function to estimate UANN relationships. The family of two F2 individuals was identified solely through UANN relationships, and therefore we incorporated the results from Sequoia along with those from COLONY 2.0.
We identified 10 pairs of FS or HS among the 15 F0individuals, indicating that the natural population was highly inbred, making it challenging to determine the detailed kinships of a reintroduced population. Despite these difficulties, we were able to identify almost all of the F2 families by combining the estimated GG, UANN, and FS relationships between F2individuals. However, despite sampling all possible grandparents for the F2 individuals analyzed, GG relationships were not estimated for all F2 individuals using COLONY 2.0. This is likely because COLONY 2.0 assumes hypothetical parents when estimating kinship (Steyaert et al., 2012), resulting in inconsistencies when attempting to estimate individuals that originally have a GG relationship by regarding them as HS.