Search

The RCA procedure is robust. A. RCA does not introduce technical variability. Each individual plasmid DNA was run as four independent replicates for both conditions (with or without RCA). The technical reproducibility of the variant frequency call was evaluated by calculating the standard deviation within the four replicates. Value are plotted for each individual position of the reference sequence, for the two plasmids sequenced. Data are directly extracted from mpileup analysis applying a VarScan p-value threshold of 0.001. Left panels: plasmid 1, right panels: plasmid 2, top panels: crude plasmid DNA (no RCA amplification), bottom panels: RCA amplified plasmid DNA. B. Origin of the nonspecific RCA products. The sequences surrounding the unique position for which a large difference was observed between the crude plasmid DNAs and the RCA products were aligned. The sequence obtained from contaminating NTC reads is shown. The common (between the two plasmids) non-accurate position is boxed (“G” in the NTC, “A” in the two plasmids). This region is part of the plasmid DNA origin of replication. (PPTX 178 kb)

Coverage drop at the position 310 human C-stretch. The human sample #12878 from the CEPH family 1463 was sequenced either following the pipeline described in this paper or from a whole genome PCR free library (generated in our lab). The relative coverage reported by mpileup was plotted against each single position of the reference genome with a zoom in the C-stretch at position 310. (PPTX 233Â kb)

Primer sequences used for qPCR. (DOCX 37Â kb)

Sanger sequencing validation for the inheritance of high level heteroplasmy SNV. Slide 1. CEPH family 1463 pedigree chart. Slide 2. Level of heteroplasmy calculated from the MitoRS data, and the Sanger sequencing data in both forward and reverse orientation. MitoRS calculated frequencies are highlighted in red (homoplasmy, > 98%) or orange (high frequency heteroplasmy, between 10% and 98%). The corresponding Sanger frequencies are highlighted in green for easier visualization. Slides 3 to 10. Chromatograms highlighting a difference between the two family members. The positions considered are shown with a black arrowhead. Slide 10. Virtual gel visualization (on a LabChip GX - Perkin Elmer) of the PCR amplification products analyzed by Sanger sequencing. (PPTX 513Â kb)

Overview of the MitoRS wet lab method. Total DNA (5 ng) is amplified by RCA. Sequencing libraries are subsequently generated using the Nextera XT kit from Illumina, starting from 1 ng amplification product. Libraries are pooled at equimolar ratios and sequenced on a HiSeq 2500 (Illumina) using rapid mode for a paired-end run of 2 x 150 cycles. (PPTX 40 kb)

The RCA procedure is robust. A. RCA does not introduce technical variability. Each individual plasmid DNA was run as four independent replicates for both conditions (with or without RCA). The technical reproducibility of the variant frequency call was evaluated by calculating the standard deviation within the four replicates. Value are plotted for each individual position of the reference sequence, for the two plasmids sequenced. Data are directly extracted from mpileup analysis applying a VarScan p-value threshold of 0.001. Left panels: plasmid 1, right panels: plasmid 2, top panels: crude plasmid DNA (no RCA amplification), bottom panels: RCA amplified plasmid DNA. B. Origin of the nonspecific RCA products. The sequences surrounding the unique position for which a large difference was observed between the crude plasmid DNAs and the RCA products were aligned. The sequence obtained from contaminating NTC reads is shown. The common (between the two plasmids) non-accurate position is boxed (“G” in the NTC, “A” in the two plasmids). This region is part of the plasmid DNA origin of replication. (PPTX 178 kb)

Overview of the MitoRS analysis methods. FastQ files are aligned with BWA to the original, and an origin-shifted version, of the DNA reference sequence (refer to Additional file 2 for details). Variant frequency is evaluated by samtools mpileup. Low frequency variants are finally identified using VarScan 2. Datasets generated from both the original reference and the shifted reference are merged, keeping only the per position values from the dataset with the highest coverage. Data are further processed to generate 1) a csv file summarizing the sequencing results observed at each individual mitochondrial DNA position and 2) a fastA file representing the corresponding consensus sequence. More details can be found in the Methods section. (PPTX 264Â kb)

RCA enriches circular versus linear templates. A. Principle of circular DNA enrichment. A single priming event will generate several concatenated copies of a circular template. At the opposite, a single copy will be amplified if the template DNA is linear. B. The RCA amplified material is mostly mtDNA. Digestion of the mouse DNA RCA product with the SpeI restriction endonuclease results in the expected mtDNA digestion product with only low amount of undigested DNA left. SpeI restriction digest is expected to result in four fragments for the B6D2F1 strain (7’398, 3’759, 3’105, and 2’035 bp) and five fragments for the NMRI strain (7’398, 3’105, 2’150, 2’037, and 1609 bp). Ladder sizes imprecision is in accordance with the Agilent Genomic DNA ScreenTape specifications. (PPTX 133 kb)

Mouse SNV frequencies are homogenous within the 88 position analyzed. A. Individual SNV frequency. For the 12 mouse mtDNA mixture ratios tested, the frequencies measured for each SNV (88 in total) were plotted and summarized as a boxplot. The boxplot whiskers highlight the extreme values (min and max). Note that the scale is different for each plot. The eight positions showing the highest frequency underestimation are highlighted in blue. These data were used to build the Fig. 4. The raw data are available from the Additional file 7. B. The eight SNV with underestimated frequency are located into two dense clusters. For each 88 SNV, the deviation from the theoretical frequency was calculated and plotted versus their position in the mitochondrial genome. The eight positions showing a systematic frequency underestimation highlighted in A. are also shown in blue. They cluster into two very short genomic regions. (PPTX 484 kb)

Overview of the MitoRS analysis methods. FastQ files are aligned with BWA to the original, and an origin-shifted version, of the DNA reference sequence (refer to Additional file 2 for details). Variant frequency is evaluated by samtools mpileup. Low frequency variants are finally identified using VarScan 2. Datasets generated from both the original reference and the shifted reference are merged, keeping only the per position values from the dataset with the highest coverage. Data are further processed to generate 1) a csv file summarizing the sequencing results observed at each individual mitochondrial DNA position and 2) a fastA file representing the corresponding consensus sequence. More details can be found in the Methods section. (PPTX 264Â kb)

Mouse SNV frequencies are homogenous within the 88 position analyzed. A. Individual SNV frequency. For the 12 mouse mtDNA mixture ratios tested, the frequencies measured for each SNV (88 in total) were plotted and summarized as a boxplot. The boxplot whiskers highlight the extreme values (min and max). Note that the scale is different for each plot. The eight positions showing the highest frequency underestimation are highlighted in blue. These data were used to build the Fig. 4. The raw data are available from the Additional file 7. B. The eight SNV with underestimated frequency are located into two dense clusters. For each 88 SNV, the deviation from the theoretical frequency was calculated and plotted versus their position in the mitochondrial genome. The eight positions showing a systematic frequency underestimation highlighted in A. are also shown in blue. They cluster into two very short genomic regions. (PPTX 484 kb)

Haplogroup determination for the CEPH family 1463. The fastA file generated for each individual of the CEPH family 1463 was submitted to the Haplofind tool. Note that N are considered as deletion. When the completion status is “No”, Haplofind was not able to determine the exact subhaplogroup. (DOCX 39 kb)

The large number of homoplasmic variants identified in the NMRI strain does not have a major impact on the mpileup reported coverage. The relative coverage reported by mpileup was plotted against each single position of the mouse reference genome for both the B6D2F1 and the NMRI datasets. The only noticeable differences are two NMRI specific “extreme” drops of coverage (positions 5’205 and 9’821) resulting from near homoplasmic indels (see the Additional file 2 for details on “extreme” coverage drops). (PPTX 264 kb)

Overview of the MitoRS wet lab method. Total DNA (5 ng) is amplified by RCA. Sequencing libraries are subsequently generated using the Nextera XT kit from Illumina, starting from 1 ng amplification product. Libraries are pooled at equimolar ratios and sequenced on a HiSeq 2500 (Illumina) using rapid mode for a paired-end run of 2 x 150 cycles. (PPTX 40 kb)

Structure of the output csv table. Example of an output file generate by the analysis pipeline. The table is populated for all positions of the reference genome. Chrom: Reference genome used, Position: Position in the reference genome, Covmp: Absolute depth of coverage, PercentCov: relative depth of coverage expressed as a percentage of the average coverage obtained for the sample, FilterCov: Flagging for insufficient relative coverage, Ref: Nucleotide from the reference genome, Var: Alternative nucleotide identified by VarScan, Cons: Consensus nucleotide kept by VarScan, FastA: Nucleotide kept in the exported fastA file, QDepth: Absolute depth of coverage, Reads1: Reference nucleotide coverage by mpileup, Reads2: Alternative nucleotide coverage by mpileup, Freq: Variant frequency, P-value: VarScan p-value, StrandFilter: VarScan strand filter, R1+: Reference nucleotide coverage from the positive strand, R1-: Reference nucleotide coverage from the negative strand, R2+: Alternative nucleotide coverage from the positive strand, R2-: Alternative nucleotide coverage from the negative strand. When several samples are analyzed together, each sample data are populated in consecutive columns of a single table. Full output tables from data presented in this manuscript can be found in the Additional files 7, 8, 9 and 13. (DOCX 46Â kb)

RCA enriches circular versus linear templates. A. Principle of circular DNA enrichment. A single priming event will generate several concatenated copies of a circular template. At the opposite, a single copy will be amplified if the template DNA is linear. B. The RCA amplified material is mostly mtDNA. Digestion of the mouse DNA RCA product with the SpeI restriction endonuclease results in the expected mtDNA digestion product with only low amount of undigested DNA left. SpeI restriction digest is expected to result in four fragments for the B6D2F1 strain (7’398, 3’759, 3’105, and 2’035 bp) and five fragments for the NMRI strain (7’398, 3’105, 2’150, 2’037, and 1609 bp). Ladder sizes imprecision is in accordance with the Agilent Genomic DNA ScreenTape specifications. (PPTX 133 kb)

Supporting Information. (DOCX 56Â kb)