Blood Ties: Chimerism Can Mask Twin Discordance in High-Throughput Sequencing

Abstract

Twin studies have long provided a means to separate the contributions of genetic and environmental factors. A recent pioneering report by Baranzini et al. presented an analysis of the complete genomes and epigenomes of a monozygotic (MZ) twin pair discordant for multiple sclerosis. This failed to find any difference between the twins, raising doubts regarding the value of whole-genome twin studies for defining disease susceptibility alleles. However, the study was carried out with DNA extracted from blood. In many cases, the hematopoietic lineages of MZ twins are chimeric due to twin-to-twin exchange of hematopoietic stem cells during embryogenesis. We therefore wondered how chimerism might impact the ability to identify genetic differences. We inferred the blood chimerism rates and profiles of more than 30 discordant twin cases from a wide variety of medical conditions. We found that the genotype compositions of the twins were highly similar. We then benchmarked the performance of SNP callers to detect discordant variations using high-throughput sequencing data. Our analysis revealed that chimerism patterns, well within the range normally observed in MZ twins, greatly reduce the sensitivity of SNP calls. This raises questions regarding any conclusions of genomic homogeneity that might be drawn from studies of blood-derived twin DNA.

Whole Genome and Exome Sequencing of Monozygotic Twins with Trisomy 21, Discordant for a Congenital Heart Defect and Epilepsy

“…Of note, DNA derived from blood may not be suitable for NGS studies of discordant twins [23]. About 70% of all monozygotic twins are monochorionic and share blood circulation in pregnancies.…”

Differences in Copy Number Variation between Discordant Monozygotic Twins as a Model for Exploring Chromosomal Mosaicism in Congenital Heart Defects

“…Moreover, in many cases, the hematopoietic lineages of MZ twins are chimeric due to twin-to-twin exchange of hematopoietic stem cells during embryogenesis [Erlich, 2011]. Therefore, CNV profiles on DNA derived from a buccal smear, fibroblasts or cardiac tissue may differ from those obtained on peripheral blood-derived DNA, as was shown by Rohrer et al in MZ twins discordant for Turner syndrome [Rohrer et al, 2004].…”

Human Embryogenesis

“…Molecular genotyping, as in forensic identification, is -more-or-less as a matter of policy -no better: when a genotyping scan shows an extra allele peak with less than 30 per cent of the strength of the main peak value, it is marked as noise and ignored. When an extra allele peak appears at 30 per cent or more of the main peak value, especially if multiple loci are involved, the sample is declared contaminated and discarded (but see Erlich, 2011). Because "everybody knows" chimerism is exquisitely rare, we do not in general look for it, and we hardly ever find what we do not believe will be there.…”

“…Faster growth of the XY cells in mixed-sex embryos might reasonably be expected to minimize the fraction of mixed-sex human chimeras that are detectably hermaphroditic. Some chimeras are discovered when twins are observed to be dizygotic (different sex is the easiest, but not the only, way to tell) and monochorionic (cf Erlich, 2011;Parva et al, 2009;Walker et al, 2007).…”