Imputation approach for deducing a complete mitogenome sequence from low-depth-coverage next-generation sequencing data: application to ancient remains from the Moon Pyramid, Mexico

Abstract

It is considered that more than 15 depths of coverage are necessary for next-generation sequencing (NGS) data to obtain reliable complete nucleotide sequences of the mitogenome. However, it is difficult to satisfy this requirement for all nucleotide positions because of problems obtaining a uniform depth of coverage for poorly preserved materials. Thus, we propose an imputation approach that allows a complete mitogenome sequence to be deduced from low-depth-coverage NGS data. We used different types of mitogenome data files as panels for imputation: a worldwide panel comprising all the major haplogroups, a worldwide panel comprising sequences belonging to the estimated haplogroup alone, a panel comprising sequences from the population most closely related to an individual under investigation, and a panel comprising sequences belonging to the estimated haplogroup from the population most closely related to an individual under investigation. The number of missing nucleotides was drastically reduced in all the panels, but the contents obtained by imputation were quite different among the panels. The efficiency of the imputation method differed according to the panels used. The missing nucleotides were most credibly imputed using sequences of the estimated haplogroup from the population most closely related to the individual under investigation as a panel.

MitoIMP: A Computational Framework for Imputation of Missing Data in Low-Coverage Human Mitochondrial Genome

“…Actually, a previous study using 1500-year-old highly degraded samples showed that a closely related population panel (haplogroup A panel) can fill more missing sites, as compared with population panels including other maternal lineages. 40 However, this previous study assessed the impact of imputation on the macro-haplogroup A lineages, but that of the worldwide maternal lineages has not been verified. In this study, we used the population panels for the worldwide mitochondrial genome lineages and also tested the effect of these panels on the imputation performance.…”

“…To ensure the robustness of the imputed alleles, our framework also introduced the parameter f , which is the threshold frequency to determine the major alleles. In this study, we followed the parameter condition ( f = 0.7, k = 5) used for kNN-based imputation procedures in Mizuno et al 40 The condition is one of the most accurate combinations of parameter values in the previous research. Our approach also uses the reference population panel based on the complete mitochondrial genome sequences.…”

Ancient DNA Studies in Pre-Columbian Mesoamerica

“…This area used to be the homeland of several Mesoamerican civilisations, including the Classic city of Teotihuacan, the Epiclassic city of Cholula, and the Postclassic Toltec, Tepanec, and Aztec Empires [ 11 ]. Some aDNA studies have focused on this region, including individuals from the Postclassic period carrying haplogroups A and B at higher frequencies ( Table 2 ) [ 48 , 50 , 51 , 52 , 53 , 54 , 55 , 56 , 57 , 58 , 63 ].…”

“…Because of poor DNA preservation, and despite using advanced high-throughput sequencing techniques, investigators had to rely on imputation methods to assign the mitochondrial subhaplogroup A2 to the individual. A major limitation of this study is the unavailability of a panel from a population both closely related to and contemporaneous with the Moon Pyramid individual to perm imputations with the highest level of confidence [ 63 ].…”

The Mitochondrial DNA Landscape of Modern Mexico

“…While some of the very first mitogenetic population studies already included individuals of Mexican descent [ 26 , 27 , 28 , 29 ], it has been repeatedly claimed since that country-wide mtDNA data from the general population was underrepresented in databases [ 13 , 18 , 20 , 30 , 31 , 32 ], together with Indigenous American genetic data in general [ 33 , 34 ]. Previous studies were either based on a small number of subjects (e.g., [ 35 , 36 ]), mtDNA segments shorter than the 1.1 kbp gold-standard control region (CR) range [ 37 ] (D-loop, nps 16024–16569 1–576; e.g., [ 16 , 18 ]), ancient DNA [ 38 , 39 ], data from geographically restricted or Indigenous population groups (e.g., [ 10 , 40 ]) where lineage representation could be heavily biased [ 41 ], US Mexicans (e.g., [ 20 , 21 ]), did not report the actual haplotypes (e.g., [ 4 , 42 ]), or a combination of these factors. The lack of general Mexico-wide mtDNA data affects both phylo- and population genetic studies aiming to reconstruct the history of human settlement and their interaction [ 20 ], as well as the forensic application of mtDNA as a vital niche marker in human identification and the assessment of biogeographic origin [ 37 ].…”