Numt Parser: Automated identification and removal of nuclear mitochondrial pseudogenes (numts) for accurate mitochondrial genome reconstruction in Panthera.

This article focuses on the development and evaluation of Numt Parser, a bioinformatic tool designed to identify and filter nuclear mitochondrial pseudogene (numt) contamination in mitochondrial DNA (mtDNA) sequencing datasets. Numt Parser classifies sequencing reads as originating from either true cytoplasmic mitochondrial (cymt) DNA or numt pseudogenes by comparing reads against reference sequences for both, enabling the removal of numt-derived reads to improve mitogenome reconstruction. Tested on whole genome shotgun data from two ancient Cape lions (Panthera leo), a genus known for numt presence, Numt Parser effectively reduced numt contamination and outperformed alternative filtering methods based on read alignment or BLAST, retaining more cymt reads and yielding higher coverage and fewer ambiguous sites. The tool requires well-characterized numt and cymt reference sequences and may be applicable to other taxa with available references, thereby enhancing the accuracy of phylogenetic and population genetic analyses involving mtDNA.