Detecting errors in mtDNA data by phylogenetic analysis

Abstract

Sequencing and documenting a sample of homologous DNA stretches is prone to copying errors in a way rather analogous to the biological replication process. Previous attempts at obtaining representative mtDNA sequences, typically of the control region, for evolutionary studies or forensic purposes have yielded rather unsatisfactory results in many cases. The key ingredient in pinpointing problems with given data is the phylogenetic analysis of closely related mtDNAs within the framework of an established worldwide phylogeny that is supported by coding region information. We develop some general rules by which likely errors in data tables can readily be detected without rereading whole sequences repeatedly. Following these guidelines, one can expect to lower the error rate by at least an order of magnitude, although it will still be hard to beat the mitochondrial gamma polymerase in precision.

Identification of Native American Founder mtDNAs Through the Analysis of Complete mtDNA Sequences: Some Caveats

“…In the absence of any partial RFLPs and HVS-II, it is difficult to decide whether a HVS-I sequence with the motif 16223-16325-16362, and additional mutations (such as one Vietnamese mtDNA from Lum et al 1998), is a member of G1a. This motif is, not unexpectedly, found twice more in 150 Japanese HVS-I sequences, but it is absent in 120 Chinese sequences of Nishimaki et al (1999), although their data have to be considered with caution in view of some likely errors (Bandelt et al 2001). …”

“…This Korean data set, however, is known to contain errors (Bandelt et al 2001). There is also confusion because a more recent study by the same investigators shows a reverse pattern for the 16362 polymorphism, so that out of nine mtDNAs with the partial motif 16223-16325 seven lack the 16362 transition, whereas the other two (P173 and H120) were evidently taken from the earlier study (Lee et al 1997).…”

Phylogeographic analysis of mtDNA variation in four ethnic populations from Yunnan Province: new data and a reappraisal

“…Errors detected in the data sets For many reasons, errors in the published HVS-I data were not infrequent, as was summarized by Bandelt et al (2001). By constructing a network relationship of the haplotypes (Bandelt et al 1995(Bandelt et al , 2000 and (near)matching them with published data sets, the potential errors in the data could be well distinguished.…”

New genetic evidence supports isolation and drift in the Ladin communities of the South Tyrolean Alps but not an ancient origin in the Middle East

“…6 -8 Substitutions at key sites were found on multiple branches of the mtDNA HVS1 network, despite being rare in other European populations. These substitutions have been described as 'phantom' mutations 6 and have been attributed to systematic sequencing errors.…”