Computational protocol: Next Generation Sequencing Mitochondrial DNA Analysis in Autism Spectrum Disorder

Similar protocols

Protocol publication

[…] We selected samples from the NIMH genetics repository ( Autism Distribution 5. All families in the repository have ASD cases diagnosed with gold standard autism diagnostic observation schedule [Lord et al., ] and autism diagnostic interview‐revised [Rutter et al., ]. Most of the 1428 families in Distribution 5 are sibships with two affected siblings. In order to increase the probability that mtDNA variants contribute to ASD, we obtained extended families that are connected through maternal lineage and sibships with four or more affected cases. Choice of extended families was based on maternal inheritance of mtDNA and we selected multi‐sib families as a part of an unpublished ASD exome sequencing project.Exomes were captured using Nimblegen SeqCap EZ Human Exome Library v2.0 (Roche, Basel, Switzerland) following manufacturer's recommendation. Exome enrichment, sequencing and base calling were performed at the University of Washington Genome Sciences Center for Mendelian Genomics as described before [Chapman et al., ]. We used MToolbox to extract mitochondrial reads, call variants and identify mitochondrial haplogroups [Calabrese et al., ). Picardtool was used for filtering out the PCR duplicates and GATK IndelRealigner was used for read realignment around indels.The hypothesis for our study was that in each family we can identify a small number of rare mtDNA VOIs that are shared between all affected cases within family. To identify VOIs we filtered out variants that define individual haplogroups and mtDNA variants with frequency ≥0.01 in 1000Genome data set (1000G) downloaded from MitoTool [Fan & Yao, ]. In addition to frequency we required VOIs to affect protein coding mtDNA (missense, frameshift, stop) and be present in all affected subjects in a family. The mtDNA VOI were confirmed with Sanger sequencing as described [Chapman et al., ]. Pathogenicity prediction and annotation for VOI were determined with MitImpact, a collection of pre‐computed pathogenicity predictions for all nucleotide changes that cause non‐synonymous substitutions in human mitochondrial protein coding genes [Castellana, Ronai, & Mazza, ].To assess interactions with nDNA in families with mtDNA VOIs, we analyzed 83 nDNA RCC genes as defined by HGNC [Gray, Yates, Seal, Wright, & Bruford, ] ( For nDNA RCC genes SNVs were annotated with ANNOVAR as described before [Rubinstein et al., ], and VOIs defined using the same variant frequency and function criteria as for mtDNA. […]

Pipeline specifications