Ancient DNA (aDNA) studies are frequently focused on the analysis of the mitochondrialDNA (mtDNA), which is much more abundant than the nuclear genome, hence canbe better retrieved from ancient remains. However, postmortem DNA damage andcontamination make the data analysis difficult because of DNA fragmentation andnucleotide alterations. In this regard, the assessment of the heteroplasmic fraction inancient mtDNA has always been considered an unachievable goal due to the complexityin distinguishing true endogenous variants from artifacts. We implemented and applieda computational pipeline for mtDNA analysis to a dataset of 30 ancient human samplesfrom an Iron Age necropolis in Polizzello (Sicily, Italy). The pipeline includes severalmodules from well-established tools for aDNA analysis and a recently released variantcaller, which was specifically conceived for mtDNA, applied for the first time to aDNAdata. Through a fine-tuned filtering on variant allele sequencing features, we wereable to accurately reconstruct nearly complete (>88%) mtDNA genome for almost allthe analyzed samples (27 out of 30), depending on the degree of preservation andthe sequencing throughput, and to get a reliable set of variants allowing haplogroupprediction. Additionally, we provide guidelines to deal with possible artifact sources,including nuclear mitochondrial sequence (NumtS) contamination, an often-neglectedissue in ancient mtDNA surveys. Potential heteroplasmy levels were also estimated,although most variants were likely homoplasmic, and validated by data simulations,proving that new sequencing technologies and software are sensitive enough to detectpartially mutated sites in ancient genomes and discriminate true variants from artifacts.A thorough functional annotation of detected and filtered mtDNA variants was alsoperformed for a comprehensive evaluation of these ancient samples.
|Numero di pagine||20|
|Rivista||Frontiers in Genetics|
|Stato di pubblicazione||Published - 2021|
All Science Journal Classification (ASJC) codes