Scientists at Eurofins in Germany have shown that they can distinguish identical twins based on single nucleotide differences in their genomes, providing a path to help solve paternity and forensic cases that involve monozygotic twins.
Current paternity and forensic genetic testing relies largely on a set of short tandem repeats, or STRs, which cannot differentiate between monozygotic twins because they are identical for both.
As such, a twin alleged to be the father of a child or the perpetrator of a crime could so far argue that the culprit might be their identical sibling.
This could change now with Eurofins' new approach, which was recently published in Forensic Science International: Genetics. The method, which relies on whole-genome sequencing, "can be used to shed light on so far unsolvable forensic paternity and criminal cases that involve identical twins," the authors wrote.
"The idea that there should be [genetic] differences in twins that could be found by next-generation sequencing is not new, but nobody had tested it so far," Burkhard Rolf, director of DNA forensic services at Eurofins Medigenomix Forensik in Ebersberg, and the senior author of the study, told In Sequence.
Following a request for testing last year from a German court regarding a paternity case that involved an identical twin father, which they had to turn down because there was no test available, Rolf and his colleagues decided to embark on a study to see whether they can actually distinguish twins by their genomes.
For their study, they recruited a pair of identical twin brothers, along with the wife and child of one of the twins.
They sequenced DNA derived from sperm samples of the twins and a blood sample of the child, using Illumina's HiSeq 2000 with 2x100 base pair reads, and generated about 90-fold coverage for each of the twins and 60-fold coverage for the child.
They then mapped the reads to the human reference genome and used the VarScan2 software to look for single nucleotide variants.
That software is designed to analyze tumor/normal pairs, discovering both germline variants and somatic mutations in the tumor. The scientists used this feature to look for variants that differ between the twins: they assigned one of the twins to be the "normal tissue" and the other to be the "tumor" and analyzed their DNA for a set of "somatic mutations." Using a similar approach, they looked for variants inherited by the child from one twin but not the other.
In total, they identified 12 SNP candidates that were present in both the twin father and the child but not the twin uncle, of which they discarded seven because their surrounding sequence could not be matched uniquely to the reference genome.
The remaining five SNPs, which were located on five different chromosomes and were not present in the dbSNP database, were confirmed by Sanger sequencing.
The researchers then tested whether the same distinguishing SNPs were not only present in sperm DNA but also in other types of DNA samples from the father. They found that four of the five SNPs could also be found in DNA from a buccal swab, but only one of the five was present in DNA from the father's blood.
This could be explained by the way these genetic differences between twins arise. Twinning occurs when then early embryo, or morula, splits into two, which according to the paper usually happens between the 16-cell and the 150-cell stage.
Mutations that distinguish the twins must occur after the morula splits, or just before that happens in a confined area. In order to be transmitted to a twin's child, they also need to be present in the father's germline.
But mutations continue to arise during the development of the embryo, so some are present in germ cells but not in other tissues, and vice versa.
Also, sperm and buccal mucosa are developmentally more closely related than sperm and blood, so some mutations will be present in both sperm cells and buccal tissue but not in blood cells, as the researchers found in their study.
Based on the study, it should be possible to identify SNPs that differ between identical twins from buccal swab DNA, a sampling method which the authors noted is "legally and ethically easier than sampling sperm" for paternity or criminal cases.
Their method for twin identification should work not only for paternity cases, they wrote, but also for forensic cases where contact stains, skin scales, hair, buccal mucosa, or semen stains were left at the crime scene.
Next-gen sequencing would only be performed on sperm or buccal samples obtained from both twins, Rolf explained. SNPs differing between them could then be detected in the crime scene samples using PCR technology, which only requires small amounts of DNA.
The method will "very likely" also work with blood stains from a crime scene, the authors wrote, though in this case, next-gen sequencing would be performed on blood samples of the twins because the blood stain might not contain the same mutations as a buccal or sperm sample.
While there might be other ways to distinguish monozygotic twins using molecular markers – for example differences in DNA methylation patterns or copy number variants – "the number of [SNP] positions we found between twins makes us optimistic that this could really become a general method to differentiate twins," Rolf said.
"We're pretty sure that we should be able to distinguish all twins, but at the moment, we have only shown it for one case," he noted.
According to Jianye Ge, an assistant professor in the department of forensics and investigative genetics at the University of North Texas Health Science Center, one problem with the approach is the high error rate of next-gen sequencing, so confirming results by Sanger sequencing will be crucial. Until a "substantial false rates study" of the technology for twin differentiation is conducted, "it won't be used in real casework," he told In Sequence.
In addition, most forensic and paternity testing labs today do not have access to next-gen sequencing equipment, Ge said, and the analysis is "pretty expensive" at the moment.
An alternative way to differentiate twins could be rapidly mutating STRs, he said, which can be assessed using standard equipment, though the accuracy would depend on the number of such STRs used.
Methylation patterns, which can be assessed by chip-based technologies, are another promising approach, he said. "Although further studies are needed … I do believe this test is more realistic for current forensic labs and less heavy-duty than NGS."
In the meantime, Eurofins is now offering its NGS method as a service for paternity testing and forensics worldwide, "and we hope that someone will charge us to do sequencing on a real case," according to Rolf.
The service has a turnaround time of approximately eight weeks, depending on the types of samples available.
Eurofins declined to disclose pricing for the service, which Rolf said would depend on the individual case, noting that sequencing costs continue to come down. High-coverage whole-genome sequencing runs into thousands of dollars per sample with other service providers.
Because all SNPs are validated by Sanger sequencing, a well-established method, Rolf expects the evidence to be accepted by the courts.
The market for twin testing could be considerable, given that about 6 in 1,000 males are identical twins. "There are many [paternity] cases in Germany that are not solved where identical twins are involved as fathers," Rolf said. Right now, "these cases end up at the [higher regional court], and the higher courts are not able to make a decision who has to pay child support."