Recovery of mitogenomes from whole genome sequences to infer maternal diversity in 1883 modern taurine and indicine cattle

Abstract

Maternal diversity based on a sub-region of mitochondrial genome or variants were commonly used to understand past demographic events in livestock. Additionally, there is growing evidence of direct association of mitochondrial genetic variants with a range of phenotypes. Therefore, this study used complete bovine mitogenomes from a large sequence database to explore the full spectrum of maternal diversity. Mitogenome diversity was evaluated among 1883 animals representing 156 globally important cattle breeds. Overall, the mitogenomes were diverse: presenting 11 major haplogroups, expanding to 1309 unique haplotypes, with nucleotide diversity 0.011 and haplotype diversity 0.999. A small proportion of African taurine (3.5%) and indicine (1.3%) haplogroups were found among the European taurine breeds and composites. The haplogrouping was largely consistent with the population structure derived from alternate clustering methods (e.g. PCA and hierarchical clustering). Further, we present evidence confirming a new indicine subgroup (I1a, 64 animals) mainly consisting of breeds originating from China and characterised by two private mutations within the I1 haplogroup. The total genetic variation was attributed mainly to within-breed variance (96.9%). The accuracy of the imputation of missing genotypes was high (99.8%) except for the relatively rare heteroplasmic genotypes, suggesting the potential for trait association studies within a breed.

Figure 1

Principal components (PC1, 2, and 3) plot based on mitochondrial genomic relationship matrix showing the grouping of I, P, Q and T major haplogroups.

Figure 2

Population structure of cattle mitochondrial sequence variants using Admixture for a pre-defined number of populations (k) ranging from 2 to 6. Population structure annotated with individual animal haplogroups (I1, I2, P, Q, T1, T2, T3, T4, T5, T6) determined from MitoToolPy.

Figure 3

Hierarchical clustering of animals based on the number of nucleotide differences between the pair of mitochondrial sequences. Cluster 1 and 2 corresponded to taurus and indicus cattle, respectively.

Figure 4

Subgrouping animals under I haplogroup into I2, I1 and subgroups within I1 (I1-Orig, I1a and I1b) using conventional Maximum Likelihood method (a) and alternate clustering techniques: principal component analysis (b), Admixture software (c) and hierarchical clustering based on the number of nucleotide difference between the sequences of pair of animals (d). *Base pair position of private variants relative to ARS-UCD1.2_M. I1-Orig is group of animals under previous I1 haplogroup not assigned to either I1a or I1b (i.e., remaining animals in I1 Cluster1).

Figure 5

Haplotype network consisting of 210 haplotypes in the Holstein population (N = 267) using the median-joining network in PopART and annotated with haplogroups predicted from MitoToolPy. The size of the circles is proportional to the number of animals carrying the same haplotype.