ReseqChip: automated integration of multiple local context probe data from the MitoChip array in mitochondrial DNA sequence assembly

The case for the continuing use of the revised Cambridge Reference Sequence (rCRS) and the standardization of notation in human mitochondrial DNA studies

Since the determination in 1981 of the sequence of the human mitochondrial DNA (mtDNA) genome, the Cambridge Reference Sequence (CRS), has been used as the reference sequence to annotate mtDNA in molecular anthropology, forensic science and medical genetics. The CRS was eventually upgraded to the revised version (rCRS) in 1999. This reference sequence is a convenient device for recording mtDNA variation, although it has often been misunderstood as a wild-type (WT) or consensus sequence by medical geneticists. Recently, there has been a proposal to replace the rCRS with the so-called Reconstructed Sapiens Reference Sequence (RSRS). Even if it had been estimated accurately, the RSRS would be a cumbersome substitute for the rCRS, as the new proposal fuses--and thus confuses--the two distinct concepts of ancestral lineage and reference point for human mtDNA. Instead, we prefer to maintain the rCRS and to report mtDNA profiles by employing the hitherto predominant circumfix style. Tree diagrams could display mutations by using either the profile notation (in conventional short forms where appropriate) or in a root-upwards way with two suffixes indicating ancestral and derived nucleotides. This would guard against misunderstandings about reporting mtDNA variation. It is therefore neither necessary nor sensible to change the present reference sequence, the rCRS, in any way. The proposed switch to RSRS would inevitably lead to notational chaos, mistakes and misinterpretations.

Inosine triphosphate pyrophosphohydrolase (ITPA) polymorphic sequence variants in adult hematological malignancy patients and possible association with mitochondrial DNA defects

Background

Inosine triphosphate pyrophosphohydrolase (ITPase) is a ‘house-cleaning’ enzyme that degrades non-canonical (‘rogue’) nucleotides. Complete deficiency is fatal in knockout mice, but a mutant polymorphism resulting in low enzyme activity with an accumulation of ITP and other non-canonical nucleotides, appears benign in humans. We hypothesised that reduced ITPase activity may cause acquired mitochondrial DNA (mtDNA) defects. Furthermore, we investigated whether accumulating mtDNA defects may then be a risk factor for cell transformation, in adult haematological malignancy (AHM).

Methods

DNA was extracted from peripheral blood and bone marrow samples. Microarray-based sequencing of mtDNA was performed on 13 AHM patients confirmed as carrying the ITPA 94C>A mutation causing low ITPase activity, and 4 AHM patients with wildtype ITPA. The frequencies of ITPA 94C>A and IVS2+21A>C polymorphisms were studied from 85 available AHM patients.

Results

ITPA 94C>A was associated with a significant increase in total heteroplasmic/homoplasmic mtDNA mutations (p<0.009) compared with wildtype ITPA, following exclusion of haplogroup variants. This suggested that low ITPase activity may induce mitochondrial abnormalities. Compared to the normal population, frequencies for the 94C>A and IVS2+21A>C mutant alleles among the AHM patients were higher for myelodyplastic syndrome (MDS) - but below significance; were approximately equivalent for chronic lymphoblastic leukemia; and were lower for acute myeloid leukemia.

Conclusions

This study invokes a new paradigm for the evolution of MDS, where nucleotide imbalances produced by defects in ‘house-cleaning’ genes may induce mitochondrial dysfunction, compromising cell integrity. It supports recent studies which point towards an important role for ITPase in cellular surveillance of rogue nucleotides.

Insights into N-calls of mitochondrial DNA sequencing using MitoChip v2.0

Background

Developments in DNA resequencing microarrays include mitochondrial DNA (mtDNA) sequencing and mutation detection. Failure by the microarray to identify a base, compared to the reference sequence, is designated an 'N-call.' This study re-examined the N-call distribution of mtDNA samples sequenced by the Affymetrix MitoChip v.2.0, based on the hypothesis that N-calls may represent insertions or deletions (indels) in mtDNA.

Findings

We analysed 16 patient mtDNA samples using MitoChip. N-calls by the proprietary GSEQ software were significantly reduced when either of the freeware on-line algorithms ResqMi or sPROFILER was utilized. With sPROFILER, this decrease in N-calls had no effect on the homoplasmic or heteroplasmic mutation levels compared to GSEQ software, but ResqMi produced a significant change in mutation load, as well as producing longer N-cell stretches. For these reasons, further analysis using ResqMi was not attempted. Conventional DNA sequencing of the longer N-calls stretches from sPROFILER revealed 7 insertions and 12 point mutations. Moreover, analysis of single-base N-calls of one mtDNA sample found 3 other point mutations.

Conclusions

Our study is the first to analyse N-calls produced from GSEQ software for the MitoChipv2.0. By narrowing the focus to longer stretches of N-calls revealed by sPROFILER, conventional sequencing was able to identify unique insertions and point mutations. Shorter N-calls also harboured point mutations, but the absence of deletions among N-calls suggests that probe confirmation affects binding and thus N-calling. This study supports the contention that the GSEQ is more capable of assigning bases when used in conjunction with sPROFILER.

Mitochondrial genome sequence analysis: a custom bioinformatics pipeline substantially improves Affymetrix MitoChip v2.0 call rate and accuracy

Background

Mitochondrial genome sequence analysis is critical to the diagnostic evaluation of mitochondrial disease. Existing methodologies differ widely in throughput, complexity, cost efficiency, and sensitivity of heteroplasmy detection. Affymetrix MitoChip v2.0, which uses a sequencing-by-genotyping technology, allows potentially accurate and high-throughput sequencing of the entire human mitochondrial genome to be completed in a cost-effective fashion. However, the relatively low call rate achieved using existing software tools has limited the wide adoption of this platform for either clinical or research applications. Here, we report the design and development of a custom bioinformatics software pipeline that achieves a much improved call rate and accuracy for the Affymetrix MitoChip v2.0 platform. We used this custom pipeline to analyze MitoChip v2.0 data from 24 DNA samples representing a broad range of tissue types (18 whole blood, 3 skeletal muscle, 3 cell lines), mutations (a 5.8 kilobase pair deletion and 6 known heteroplasmic mutations), and haplogroup origins. All results were compared to those obtained by at least one other mitochondrial DNA sequence analysis method, including Sanger sequencing, denaturing HPLC-based heteroduplex analysis, and/or the Illumina Genome Analyzer II next generation sequencing platform.

Results

An average call rate of 99.75% was achieved across all samples with our custom pipeline. Comparison of calls for 15 samples characterized previously by Sanger sequencing revealed a total of 29 discordant calls, which translates to an estimated 0.012% for the base call error rate. We successfully identified 4 known heteroplasmic mutations and 24 other potential heteroplasmic mutations across 20 samples that passed quality control.

Conclusions

Affymetrix MitoChip v2.0 analysis using our optimized MitoChip Filtering Protocol (MFP) bioinformatics pipeline now offers the high sensitivity and accuracy needed for reliable, high-throughput and cost-efficient whole mitochondrial genome sequencing. This approach provides a viable alternative of potential utility for both clinical diagnostic and research applications to traditional Sanger and other emerging sequencing technologies for whole mitochondrial genome analysis.

Large scale mtDNA sequencing reveals sequence and functional conservation as major determinants of homoplasmic mtDNA variant distribution

The mitochondrial DNA (mtDNA) is highly variable, containing large numbers of pathogenic mutations and neutral polymorphisms. The spectrum of homoplasmic mtDNA variation was characterized in 730 subjects and compared with known pathogenic sites. The frequency and distribution of variants in protein coding genes were inversely correlated with conservation at the amino acid level. Analysis of tRNA secondary structures indicated a preference of variants for the loops and some acceptor stem positions. This comprehensive overview of mtDNA variants distinguishes between regions and positions which are likely not critical, mainly conserved regions with pathogenic mutations and essential regions containing no mutations at all.

Searching for a needle in the haystack: comparing six methods to evaluate heteroplasmy in difficult sequence context

Mitochondrial DNA (mtDNA) mutations have been involved in disease, aging and cancer and furthermore exploited for evolutionary and forensic investigation. When investigating mtDNA mutations the peculiar aspects of mitochondrial genetics, such as heteroplasmy and threshold effect, require suitable approaches which must be sensitive enough to detect low-level heteroplasmy and, precise enough to quantify the exact mutational load. In order to establish the optimal approach for the evaluation of heteroplasmy, six methods were experimentally compared for their capacity to reveal and quantify mtDNA variants. Drawbacks and advantages of cloning, Fluorescent PCR (F-PCR), denaturing High Performance Liquid Chromatography (dHPLC), quantitative Real-Time PCR (qRTPCR), High Resolution Melting (HRM) and 454 pyrosequencing were determined. In particular, detection and quantification of a mutation in a difficult sequence context were investigated, through analysis of an insertion in a homopolymeric stretch (m.3571insC).