Sequencing of maternal plasma to detect fetal chromosomal aneuploidies is currently being offered by four US-based companies — Sequenom, Verinata Health, Ariosa Diagnostics, and Natera — and researchers now think that the technique can be expanded to detect sub-chromosomal deletions and duplications in the fetal genome as well.
Such aberrations are less frequent than fetal aneuploidy, but can cause conditions like DiGeorge syndrome, characterized by congenital heart defects, learning deficiencies, immune problems, abnormal facial features, and other problems.
Two separate groups — from Illumina's Verinata Health and Dennis Lo's group at the Chinese University of Hong Kong, from whom Sequenom licenses IP for its MaterniT21 Plus aneuploidy test — recently published proof-of-principle studies on methods to detect sub-chromosomal abnormalities in fetal DNA and are now looking to validate their methods in larger cohorts and eventually incorporate them into a commercial test.
Researchers at Verinata developed a method that uses shotgun sequencing and then partitions the sequenced fragments into 1-megabase-sized bins to determine over- or under-representation on any chromosomes.
Lo's team's approach is very similar, except for slight differences in the sequencing technique and the researchers also sequenced to less depth in the study, so the method had a slightly lower resolution.
Verinata CSO Rich Rava, told Clinical Sequencing News that the company is now working out how to incorporate such sub-chromosomal conditions into a commercial test.
The firm is currently deciding which specific sub-chromosomal deletions and duplications to include and validating the method for those abnormalities, as well as determining how to offer the test — whether to include it into the aneuploidy test or only offer it as a reflex test if there is an abnormal result on an ultrasound.
Likewise, Lo said that ideally all high-risk pregnancies would be offered such testing included with the aneuploidy test after it is further validated. However, he anticipated that the method would add additional cost since it requires more sequencing, so "one option is to offer the noninvasive molecular karyotyping test to fetuses with significant ultrasound abnormalities."
Lo's group demonstrated its sequencing-based karyotyping technique on six cases with known microdeletions and microduplications and eight known normal cases.
The technique, published in PLoS One in April, identified all known aberrations at a higher resolution than standard chromosomal karyotyping, and Lo said that higher resolution can be achieved by greater sequencing depth.
Using sequencing to detect fetal aneuplodies requires sequencing to a low depth of coverage, typically around 0.5x coverage of the genome. To detect sub-chromosomal aberrations, more sequencing is required. In the PLoS One study, the researchers sought to obtain 99 percent sensitivity for detecting structural changes down to 3 megabases in size. To do this, they used 50-base paired end sequencing on one lane of a flow cell on an Illumina HiSeq 2000, which generated an average of 144 million aligned reads per sample, corresponding to 4.81-fold haploid coverage.
Then, to identify the structural changes, they used a binning approach, dividing the chromosomes into 2,687 1-megabase bins, calculating the genomic representation of each bin.
To call copy number variations, Lo's group determined a z-score for each bin — the same metric used to call fetal aneuploidy. First, the researchers used eight normal samples to determine a z-score and standard deviation for each bin.
When assessing the test cases, normal regions of the genome would be expected to have a score close to zero. The reference interval was defined as a z-score between -3 and +3. To reduce false positives, the researchers only called a copy number aberration if three consecutive 1-mb bins exhibited a z-score outside the reference interval.
From the six test cases, the researchers confirmed de novo microdeletions in three cases, one de novo microduplication, one inherited microduplication, and one case in which the mother was a carrier for a balanced translocation and the fetus had both a microduplication and deletion. The team had no false positives.
For all the cases, the sizes of deletions and duplications corresponded to what was detected by other techniques, such as array-CGH, fluorescence in situ hybridization, and quantitative fluorescence PCR.
The group then used computer simulations to determine the diagnostic sensitivity of their approach, finding that the sensitivity for detecting a 3-megabase aberration would be around 96 percent with a fetal DNA percentage of 5 percent, but could increase to 99 percent with 6 percent fetal DNA.
To achieve 99 percent diagnostic sensitivity at a 1-megabase resolution, much more sequencing is needed — around 480 million aligned reads must be generated, corresponding to nearly 15-fold coverage of the genome.
Verinata's method is very similar. The company presented a poster on it at last year's American Society of Human Genetics meeting in San Francisco (CSN 11/14/2012), and published a proof of principle in the American Journal of Human Genetics.
Since the techniques are very similar — both rely on whole-genome shotgun sequencing and then dividing the genome into 1-megabase sized bins — "our study has therefore independently validated the feasibility of performing prenatal whole-genome molecular karyotyping noninvasively," Lo said.
There are slight differences in their shotgun sequencing approach, however. Verinata used single-end 25-base sequencing, while Lo's team used 50-base paired end reads. Additionally, the Verinata team generated around 960 million reads per sample, compared to the 144 million reads from Lo's team.
Rava added that the differences relate to "what resolution can you use to look at copy number variants across the genome," he said, with more sequencing yielding greater resolution.
In its initial study, Verinata tested the method on samples from 11 women with fetuses with sub-chromosomal deletions and duplications, translocations, mosaicism, and a case of trisomy 20 that had all been diagnosed with metaphase karyotyping.
The protocol was able to identify the fetal deletions, duplications, translocations, and trisomy 20, but not the four cases of mosaicism.
Similar to Lo's method, the Verinata team partitioned the sequenced fragments into 1-megabase bins to determine over- and under-representation on specific chromosomes.
Consistent with the fetal metaphase karyotypes, the team identified a 38-megabase duplication on chromosome six, a 300-kilobase deletion on chromosome 7 in two separate samples, a 6-megabase deletion on chromosome 8, a 17-megabase deletion on chromosome 15, a 19-megabase duplication to chromosome 17, a 2-megabase deletion on chromosome 10, a 40-megabase duplication on chromosome 3, and a 9-megabase deletion on the X chromosome.
Additionally, the sequencing technique was able to identify aberrations that were not picked up by metaphase karyotyping. For instance, one sample had an unbalanced translocation, while another had a 300-kilobase duplication on chromosome 7. These additional aberrations were all smaller than 1-megabase in size, indicating that the sequencing technique may have greater resolution than metaphase karyotyping, a result consistent with Lo's team's findings.
The next step for both groups will be deciding which sub-chromosomal aberrations to include in a test and to validate the analytical sensitivity and specificity for such indications.
While each individual sub-chromosomal abnormality is rare, collectively, around "4 percent of high-risk pregnancies have pathological or potentially clinically significant chromosomal structural abnormalities that are not the common whole chromosome aneuploidies or sex chromosome aneuploidies," Lo said.
He said incorporating such testing into high-risk pregnancies could help detect serious fetal conditions.
For instance, the three de novo microdeletions were 22q11.2 deletions, indicative of DiGeorge syndrome, one of the more common sub-chromosomal aberrations, affecting around one in 4,000 births.
Individuals with DiGeorge syndrome have variable presentations, including cardiac malformations, immune deficiencies, developmental delay, renal problems, hypoparathyroidism, hearing deficits, and other clinical problems, Lo told CSN.
For instance, in his study, the three cases with the 22q11.2 microdeletion showed signs of heart abnormalities on the ultrasound. Ultrasound also identified thick nuchal translucency on the fetus with a de novo 22q11.2 microduplication.
The incidence of 22q11.2 microduplication is around half that of the deletion syndrome, Lo said, and the syndromes tend to be milder, but include developmental delay, hearing deficits, and cardiac abnormalities.
While the 22q11.2 deletion and duplication syndromes are known syndromes and can often be severe, cases are still highly variable. Most instances are de novo, but one fetus was found to have an inherited form of the 22q11.2 microduplication, but the mother had no symptoms.
Despite this clinical variability, Verinata's Rava said there are some common symptoms. For instance, he said around 40 percent of individuals with DiGeorge syndrome tend to have congenital heart abnormalities, and if the condition is diagnosed prenatally, then accommodations can be made to start treatment immediately after birth, he said.
He said Verinata would incorporate into a test only "things where there's something that's actionable by the doctor or patient."
He said Verinata is currently evaluating two different ways in which to incorporate subchromosomal testing: either as a reflex test or included in the aneuploidy test.
Currently, he said the company is talking with Ob-Gyns, maternal and fetal medicine specialists, and genetic counselors to determine the community's preference. "That's still a debate," he said. "Some doctors want the results when they get the aneuploidy test," he said. But many doctors prefer to do the tests separately, only looking for sub-chromosomal abnormalities if there is an unusual finding on the ultrasound, for instance.
At least initially, Verinata will likely offer it as a reflex in the case of abnormal findings elsewhere, he said, but "that could evolve."
Rava said the company has not yet determined when it would begin offering such testing. The firm is still going through the process of determining which conditions to include and validating the analytical capabilities for each condition, which will take longer than the process for trisomy 21, since these sub-chromosomal conditions are rarer.
Additionally, he thought the cost of such a test would be comparable to microarray-based invasive testing at around $1,000.