Estimating chimerism levels in marmosets through NGS analysis.
To estimate genetic diversity and devise strategies by which it can be enhanced in common marmosets that were born and maintained at the Non-Clinical Evaluation Center of KBIO Health®, and in their ancestors that were imported from CLEA Japan, Inc. (Tokyo, Japan), we constructed 62 NGS libraries from 56 marmoset specimens (Supplementary Table S1). The majority of these samples were derived from hair follicles (56), with additional samples derived from blood (5) and cultured fibroblasts (1). One sample, I4938, provided three tissues, including hair follicles, blood, and cultured fibroblasts. In total, we produced 35 Gbp NGS reads with an average coverage of 39× per individual. We also incorporated 31 Callitrichidae NGS data for comparison from the National Center Biotechnology Information (NCBI) database, with genomic coverage ranging from 11.65× to 55.9× (Supplementary Table S2). We confirmed marmoset chimerism in NGS data through comparison with the distribution of allelic balance observed in chimeric-free human data (Fig. 1a). Marmoset data showed a broad distribution of allelic balance values near homo-alleles that suggests the presence of chimerism from twin sibling cells. In sample I4938, we observed varying levels of chimerism across tissues, with these variations being more pronounced in the blood samples (Fig. 1b and Supplementary Fig. S1). We calculated the chimerism ratio in twin samples (1722300M and 1722301M) based on the peak of allelic balance in 4.8M SNVs in the “common marmoset SNV panel”. As a result, we observed that the peaks of allelic balance were twice as high when the sibling's genotype was homozygous (peak of the allelic balance = 0.032 for 1722300M) compared to when there were heterozygous genotypes (0.016) (Fig. 1b). This tendency remained unchanged both when calculated in ref-homo regions (0.98 vs. 0.987) of 1722300M and when calculated for the other twin individual of 1722301M (Supplementary Fig. S2). We observed relatively low chimeric ratios in hair follicles, with an average of 5% and a range from 2–9%, with increasing average values of 6% found in ear punch and 25% found in tail biopsy (Table S3). However, we excluded the blood samples from further analysis due to the high chimerism found therein (Supplementary Fig. S3).
Construction of Callitrichidae variome to study genetic diversity
The genome-wide allelic balance distribution was predominantly skewed toward one, which indicates the proportion of reference homozygous (ref-homo) alleles (Fig. 2a). Further, by analyzing the lower 1M data points, we were able to differentiate homozygous from heterozygous alleles, finding that heterozygous alleles were most frequent within an allelic balance range of 0.2–0.8 (Fig. 2b). This SNV stratification substantially reduced the misclassification of homozygous alleles stemming from chimerism, thereby enhancing the precision of SNV genotyping. Individually, we corrected 7.5% of heterozygous SNVs impacted by chimerism to more reflect their homozygous alleles and designated 24% of loci as missing due to insufficient read depth (Supplementary Table S4). After these corrections, we identified approximately 4.2 million SNVs for C. jacchus, with a range from 3.6 to 5.9 M. We identified 14 M SNVs for C. geoffroyi and 9.2 M SNVs for C. kuhlii. As we were preparing the manuscript, NGS data for 12 species across six Callitrichidae genera were released14. From this data set, we additionally identified an average of 51 M SNVs, ranging from 35 to 88 M, which were found to be contingent upon the genetic distance from C. jacchus reference.
Genetic diversity and phylogenetic relationships within the Callitrichidae family
Our study on the genetic diversity of the Callitrichidae family, which was particularly focused on the KBIO Health® C. jacchus population, revealed signs of inbreeding as well as enhanced genetic diversity enhancement as a result of inter-species admixture. Our precision genotyping highlighted fine-scale genetic differences that aligned with existing phylogenetic trees14. The analysis identified a distinct clade of 62 C. jacchus individuals, with 52 from KBIO Health® and the rest from external sources (Fig. 3a), and it found that C. kuhlli is more closely related to C. jacchus than C. geoffroyi, with Mico as a sister clade. This phylogenetic arrangement makes it easier to compare genetic diversity within the family. Moreover, a correlation was observed between genetic diversity (measured by median pi, π) and conservation status (IUCN classification), with notable variations across genera (Fig. 3b). For instance, Leontopithecus, which is categorized as 'Endangered', had the lowest π value (median π = 0.00082), while C. geoffroyi, which is categorized as ‘Least Concern’, had a higher π value (median π = 0.00308). Other anomalies included the finding of a lower π value in 'Least Concern' Mico (median π = 0.00084), which is indicative of high inbreeding and therefore reduced genetic diversity, compared to 'Near Threatened' Callibella humilis (median π = 0.00152). A similar situation was observed in a parallel manner in the marmoset population of KBIO Health®, which come from a small founding population (π = 0.0005). By stark contrast, the individual labeled SRS7106849 from the C. jacchus species exhibited two times higher genetic diversity (π = 0.00098) than other C. jacchus individuals, thus showing the significant variation in genetic diversity within the species. In the structure analysis, the K = 3 result illustrated the fundamental genetic structure of Callitrichidae phylogeny, and three Callithrix species were found to be genetically separated at K = 7. Notably, the SRS7106849 individual from C. jacchus, which exhibited twice the genetic diversity compared to other individuals within the species, was significantly influenced by genetic admixture with C. kuhlii, thus highlighting the importance of admixture in enhancing genetic diversity. This finding suggests that genetic admixture within C. jacchus and other Callitrix species can increase genetic diversity in nature. This insight is potentially valuable for conservation efforts and the development of specific traits across the species.
Inter-species genetic admixture and its impact on marmoset genetic diversity
In the present study, we aimed to uncover the genetic clustering and diversity within the KBIO Health® marmosets, starting with limited knowledge of the founder C. jacchus population from Japan. The PCA analysis results suggest the existence of high inbreeding within the KBIO Health® marmosets and the Japanese founder population. We identified three distinct genetic groups. One standout individual, SRS7106849 in Group (G1), showed significant genetic differences, including admixture with C. kuhlli, which set it apart from the other two groups (Fig. 4a). The Fst values reveal significant genetic differentiation between this unique individual and the other two groups, with Fst=0.33 for G2 and Fst=0.26 for G3, thus highlighting notable genetic stratification. Most of the individuals, including three deposited in NCBI from Japan, were grouped in Group G3, suggesting possible inbreeding within both the Japanese and our populations. By contrast, five individuals, including three from KBIO Health®, formed a separate Group G2. Pedigree analysis identified the three G2 individuals as part of the I5528 family. The genetic differentiation between G2 and G3 was relatively small (Fst=0.14), indicating closer relatedness compared to their differentiation from G1. Notably, the G2 individuals exhibited higher genetic diversity (π = 0.00064) than those in G3 (π = 0.0005) (Fig. 4b), suggesting a varied genetic background. Most remaining individuals, including the three deposited in NCBI from the Japanese population, were clustered in Group G3, indicating potential inbreeding among the Japanese population and our populations. Five others, with three from KBIO Health®, formed a distinct Group G2. Pedigree analysis showed that the three in G2 are part of the I5528 family. The genetic differentiation between G2 and G3 (Fst=0.14) indicated that those two groups were closer compared to G1. G2 individuals exhibited higher genetic diversity levels (π = 0.00064) compared to G3 (π = 0.0005) (Fig. 4b). We confirmed the unique genetic makeup of SRS7106849 using advanced genetics statistics. This individual was genetically closer to C. kuhlli than to other C. jacchus in the study, suggesting the existence of a mix of genetic components from different species (Fig. 4c). Notably, the ancestor of the individual SRS7106849 showed evidence of admixture between C. kuhlli and C. jacchus (Fig. 4d); however, no such admixture with C. kuhlli was evident in individuals within G2 and G3 (Fig. 4e). These findings emphasize the importance of maintaining a diverse founder population for preclinical research populations. These results are expected to help inform further international collaborative research in marmoset diversity and conservation strategies.