Genome, epigenome and RNA sequences of monozygotic twins discordant for multiple sclerosis

Abstract

Monozygotic (MZ) or “identical” twins have been widely studied to dissect the relative contributions of genetics and environment in human diseases. In multiple sclerosis (MS), an autoimmune demyelinating disease and common cause of neurodegeneration and disability in young adults, disease discordance in MZ twins has been interpreted to indicate environmental importance in its pathogenesis1–8. However, genetic and epigenetic differences between MZ twins have been described, challenging the accepted experimental paradigm in disambiguating effects of nature and nurture.9–12 Here, we report the genome sequences of one MS-discordant MZ twin pair and messenger RNA (mRNA) transcriptome and epigenome sequences of CD4+ lymphocytes from three MS-discordant, MZ twin pairs. No reproducible differences were detected between co-twins among ~3.6 million single nucleotide polymorphisms (SNPs) or ~0.2 million insertion-deletion polymorphisms (indels). Nor were any reproducible differences observed between siblings of the three twin pairs in HLA haplotypes, confirmed MS-susceptibility SNPs, copy number variations, mRNA and genomic SNP and indel genotypes, or expression of ~19,000 genes in CD4+ T cells. Only two to 176 differences in methylation of ~2 million CpG dinucleotides were detected between siblings of the three twin pairs, in contrast to ~800 methylation differences between T cells of unrelated individuals and several thousand differences between tissues or normal and cancerous tissues. In the first systematic effort to estimate sequence variation among MZ co-twins, we did not find evidence for genetic, epigenetic or transcriptome differences that explained disease discordance. These are the first female, twin and autoimmune disease individual genome sequences reported.

Sex differences in DNA methylation and expression in zebrafish brain: a test of an extended ‘male sex drive’ hypothesis

“…Methylation calls from RRBS technology has been shown to be very reproducible by several groups across the world [46][47][48][49][50][51][52][53]. Further, in the identified DMC cluster we see multiple adjacent CpG sites exhibits large methylation changes in the same direction.…”

Optimized filtering reduces the error rate in detecting genomic variants by short-read sequencing

“…A previous whole-genome sequencing experiment of monozygotic twins discordant for multiple sclerosis did not, however, identify true genetic differences 30 .…”

“…Another commonly used approach is to apply quality filters that are aimed at selectively removing errors. Every whole-genome sequence reported so far has used filtering to some extent: the most commonly used filters being those that remove sequences with a too-low coverage depth, discard variants with a low-confidence score or eliminate variants located within a cluster of variants 3,7,[10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25] . Surprisingly, there is little consensus with respect to which filters should be used and at which threshold they should be applied.…”

“…As a result, each reported study developed its own heterogeneous set of filters and applied them at various (suboptimal) thresholds. For instance, in the case of the coverage depth filter, thresholds removing sequences with a coverage depth <4×, <10× or <11× were applied 6,10,11,16,17,23,24 . Additionally, it has not been assessed to which extent filters discard true variants, and how each filter can be optimized in terms of sensitivity and specificity.…”

Multiple sclerosis genetics

“…Preliminary studies suggest that every individual's DNA contains over 50 000 SNP variants and over 5000 insertion/deletion polymorphisms (Baranzini et al, 2010). Importantly, 42% of the SNPs and 86% of the insertion/deletion polymorphisms are novel, meaning that they had not been previously recorded in publicly available databases.…”