By Monica Heger
This article has been updated from a version posted Dec. 9 to include additional comments from Dennis Lo as well as outside comment.
For companies looking to develop noninvasive prenatal diagnostic tests based on circulating fetal DNA, distinguishing between the maternal and fetal genomes has been a major hurdle. While several companies have been moving ahead with noninvasive sequencing-based tests for diseases caused by aneuploidy, such as trisomy 21, diagnosing more complex genetic disorders has proven to be more challenging.
Overcoming one hurdle, researchers from the Chinese University of Hong Kong and Sequenom have used Illumina sequencing to show that the entire fetal genome is present in maternal plasma — an important step toward demonstrating that "a noninvasive genome-wide scan of the fetal genome from maternal plasma is possible," the researchers wrote in a paper describing the study last week in Science Translational Medicine.
In the proof-of-concept paper, the authors also outline a method to assemble a fetal genomic map based on Illumina reads, which they used in a clinical setting to diagnose a fetus as a carrier of beta-thalassemia, a type of genetic anemia common in Southeast Asia.
Dennis Lo, professor of chemical pathology at the Chinese University of Hong Kong and lead author on the paper, told In Sequence that being able to show that the entire fetal genome is present in the maternal plasma is important because "then one might be able to scan the entire fetal genome for genetic disorders using this noninvasive approach."
The results indicate that "with further reductions in the error rate of massively parallel sequencing platforms, it may be possible that de novo mutations in the fetal genome could be detected in a cost-effective manner by maternal plasma DNA sequencing," the authors wrote.
Lo has also collaborated with Sequenom to develop its sequencing-based Down syndrome test, which the company has said will be available in the fourth quarter of 2011 (IS 9/21/2010). Additionally, Stanford University's Stephen Quake has also described a sequencing-based approach to detect Down syndrome and has licensed the IP co-exclusively to Fluidigm and Artemis Health. LifeCodexx of Germany is also working on developing non-invasive prenatal diagnostic tests based on sequencing, but has not disclosed details on its method, or for which diseases it would test (IS 5/4/2010).
LifeCodexx's CEO Michael Lutz told In Sequence that while this new approach "is scientifically very interesting and may set the basis for more analysis," it is also "very time consuming and very expensive … So, I'm not sure how you could apply this in a commercial setting."
Ian Clements, Sequenom's senior director of investor relations and corporate communications, said it was premature to say whether the company would license this technology from Lo. "Clearly we are looking to position ourselves as a leader in the noninvasive prenatal diagnostics market…so any IP or technology that could help build that over whatever period of time is something we'll be interested in talking about," he told In Sequence. But, he noted, it's still "very early stage."
Clements added that Sequenom is still on target to launch its trisomy 21 test in the fourth quarter of 2011.
What makes the Science Translational Medicine study different from the methods used to detect Down syndrome is that it is "much more complex than the use of massively parallel sequencing for the detection of fetal chromosomal aneuploidies," the authors note.
Unlike trisomy 21 testing, which only requires the detection of an extra chromosome 21, diagnosing other genetic diseases such as beta-thalassemia requires the detection of single nucleotide variations in specific areas of the fetal genome within the context of the maternal genome.
In their study, the researchers first did paired-end sequencing of maternal plasma DNA from the woman of a couple attending an obstetrics clinic for the prenatal diagnosis of beta-thalassemia. They used the Illumina Genome Analyzer, obtained read lengths of 50 base pairs, and sequenced to an average of 65-fold coverage, generating about 3.9 billion reads.
The team then used genotype data from the parents in conjunction with the sequence data to parse the fetal genome, which comprises only around 10 percent of the total plasma DNA.
First, the team searched for the portion of the fetal genome from the father's genotype. They compared maternal and paternal genetic maps, identifying markers present only in the paternal genome. The presence of paternal markers in the plasma DNA indicated that it was from the fetal genome.
For the half of the fetal genome that was inherited from the mother, the team developed an approach called relative haplotype dosage analysis, which involves looking at the ratio of DNA in the plasma that came from the mother's mother versus the mother's father. For the proof-of-concept study, the authors actually derived the maternal haplotype information from microarray analysis of a chorionic villus sample of the fetus, but they note that "in real diagnostic scenarios, the latter information would not be available."
As a result, the maternal haplotype would need to be deduced either by genotype information from other family members; by probabilistic deduction using known haplotype information within a population; "by other methods that would allow direct haplotype information to be generated," such as single-molecule analysis; or, eventually, by whole-genome sequencing of both parents.
In order to test for beta-thalassemia, the researchers looked specifically at the HBB gene. Both the mother and the father were heterozygous for a mutation in the gene, but the fetus would have had to inherit both mutations to have the disease. Analyzing the paternal genome showed that the fetus inherited the mutant copy from the father, but analysis of the maternal genome sequences showed that the mother had passed on the wild-type copy, making the fetus a heterozygous carrier of the disease.
Because sequencing the entire fetal genome is so much more complex than detecting an extra chromosome, the cost is much higher. Lo estimated that sequencing the fetal genome to 65-fold coverage, as he did in this study, would run nearly $200,000. However, he said one way to lower the cost would be to take a more targeted approach to sequencing, which could reduce the cost to $1,000 to $2,000.
Taking a targeted sequencing approach would also help improve the sensitivity and specificity of the method, necessary steps to make it clinically relevant.
Even so, the study has "opened up the possibility that one can screen for multiple genetic disorders that are prevalent in a particular population with a single test," Lo said. For example, he said, one test could screen for various genetic anemias like beta-thalassemia that are common in Southeast Asia, or in Africa a test to screen for sickle-cell anemia, and in Caucasian populations to screen for cystic fibrosis.
Additionally, the method could have other applications, such as for identifying circulating tumor DNA or for monitoring organ rejection in transplant patients, Lo said. The basic idea is the same for these applications as for detecting fetal DNA, he said. "In both you basically have an alien source of DNA that is floating in the plasma."
Lo said he has applied for a patent on the technique and is also exploring commercial options, but declined to be more specific about which companies he is talking to.
Have topics you'd like to see covered in In Sequence? Email the editor at mheger [at] genomeweb [.] com.