This article has been updated to clarify the turnaround time for Natera's test. It is two weeks.
Natera has released data showing that its approach to non-invasive, sequencing-based fetal aneuploidy testing can detect trisomies 21, 18, and 13, as well as X and Y chromosome aneuploidies, with close to 100 percent accuracy.
In a proof-of-principle study that appeared last month in Prenatal Diagnosis, Natera said that its assay uses SNP-based targeted sequencing and Bayesian statistical analysis to compare possible fetal genotypes to actual sequencing data to predict a particular aneuploidy. According to the study, the method correctly reported chromosome copy number for all five target chromosomes — 21, 18, 13, X, and Y — yielding 725 out of 725 correct calls in 145 samples that passed a DNA quality test out of a total of 166 sampled.
Following the publication of the study, at last week's annual meeting of the American Society of Human Genetics Natera released additional results from an ongoing clinical trial using a new version of its algorithm and a larger group of SNP targets — 19,500 versus the 11,000 used in the Prenatal Diagnosis study.
At the meeting, Natera presented a poster with results from 407 maternal samples analyzed using the new algorithm, which the company calls "NATUS," or "Next-generation Aneuploidy Testing Using SNPs." The updated approach demonstrated 100 percent specificity and sensitivity when detecting aneuploidies including trisomies 21, 18, and 13, and sex chromosome abnormalities, such as monosomy X, the company reported.
"In the study, we got all of our calls right," Natera's chief technology officer Jonathan Sheena told Clinical Sequencing News. "We had no errors, no false positives, no false negatives. But then … with increasing our assay from about an 11,000-plex PCR reaction to almost 20,000-plex, that really helps us keep our sensitivity and specificity up where they are and it also helped us reduce our no-call rate down to well within the rest of the industry," he said.
Sheena said Natera plans to launch the test commercially by the end of this year.
The company announced its intentions in the non-invasive prenatal sequencing space in January (CSN 1/18/2012), but initially released few details about its sequencing approach or how it differed from several other companies either currently offering, or planning to offer, non-invasive prenatal tests.
In January, Natera preemptively sued Sequenom, claiming that its method does not infringe Sequenom's patent on noninvasive prenatal diagnosis and that the patent is invalid.
Sequenom has in turn sued Natera, as well as two other companies, Ariosa and Verinata, for infringing the patent at the heart of its MaterniT21 test, which has been on the market for more than a year.
Natera's recent paper illustrates some of the ways its own sequencing approach differs from these potential competitors. Like other methods, Natera's test starts with a sample of maternal blood, which contains cell-free fetal DNA in addition to maternal DNA.
Sheena noted that most current methods then go on to quantify the amount of chromosomal DNA, comparing the sequence data to a reference chromosome in order to establish the presence of a fetal aneuploidy in the sample.
Natera's method, on the other hand, uses PCR to amplify a panel of target SNPs that are then sequenced. Sheena said the company is currently using the Illumina HiSeq 2000 for sequencing, but is platform agnostic in the long term.
The company's algorithm then uses parental genotype and other reference data to create billions of virtual fetal genotypes, predicting what possible monosomies, disomies, and trisomies would look like in sequencing data from samples containing different fetal cell-free DNA concentrations.
Using Bayesian statistics, Natera then calculates the relative likelihood of each of these hypothetical genotypes given the actual sequencing data.
According to Sheena, this approach allows the company to obtain accurate results not only for specific chromosomes, but also across a wider range of chromosomes than other tests have been able to analyze.
"We are looking at SNPs and comparing that to a clean maternal genomic sample. Using that, we can get consistently accurate results across every chromosome … about two and a half to three times the clinical coverage of anybody else's test," he said.
"What everybody else is doing is some version of counting, looking at the quantitative signal that comes off their sequencing assays and trying to calibrate how much of chromosome 21 signal do they see and compare that against a reference chromosome. [With that approach] the differences between a trisomy and a disomy are going to be very small, given that you may have only a few percent fetal [DNA]."
"Because different chromosomes amplify at different rates and there is variance in that amplification, you'll get noise, so no matter how good sequencing platforms get, the counting method is subject to amplification biases," Sheena said.
"But because we are looking at SNPs, we have much more information … to tell what the maternal contribution and paternal contributions look like. From that we can actually normalize away those amplification variations," he explained.
According to Sheena, this gives Natera a leg up in covering a wider group of chromosomal abnormalities. In the recent paper, the test identified not just trisomies in chromosomes 21, 18, and 13, but also 47,XXY and 45,X samples.
All together, the assay correctly identified 706 euploid chromosomes and 19 aneuploid chromosomes: eleven trisomy 21; three trisomy 18; two trisomy 13; one 45,X; two 47,XXY; 57 46,XX; and 69 46,XY samples.
According to the study authors, based on DNA mixing experiments and prior assay performance, Natera expects that the test will also be able to detect monosomy 21, 47,XXX and 47,XYY, and uniparental disomy and triploidy.
Sheena said that Natera's Bayesian prediction approach uniquely allows for sample-specific accuracy calculations. Overall, for the 145 samples the assay was able to call in the study, the company calculated a mean accuracy of 99.92 percent.
Natera's method uses a data quality test so that any sample with less than a four percent fetal DNA fraction or with a low DNA quality based on other metrics like noise levels is reported as a no-call.
In the study, the 21 samples (20 putative euploid samples and one 45,X) that could not be called all clustered in the low fetal DNA fraction range, according to Natera. Overall, the study demonstrated a 12.6 percent no-call rate.
Natera highlighted, however, the fact that while its test may suffer more no-calls at low fetal DNA fractions, massively parallel shotgun sequencing methods like Sequenom's and others' tend to make incorrect calls at similarly low levels. "MPSS tends to make incorrect calls on low quality samples, whereas the [Natera] method tends to make no-calls. The presumption is that a no-call is preferred to a false negative result, as a no-call simply requires a redraw and retest, whereas a miscall can result in lifelong consequences," the authors wrote.
In the more recent data presented at ASHG and other meetings, Natera has reported a much lower no-call rate than in the initial study: about six percent.
Sheena also said that if the possibility of a second sampling and test procedure is included in calculations, the overall no-call rate should go down even further. According to the company, testing currently has a turnaround time of about two weeks, which should allow for quick enough retesting to avoid reflex to invasive procedures after an initial no-call.
Natera currently offers a non-invasive paternity test based on the same sequencing approach, which it has studied for secondary analysis result rates. Sheena said the company has calculated that when initial paternity no-calls have gone on to submit a second sample, about 95 percent of those have yielded a result the second time around. That comes down to about a quarter of a percent no-call rate after a single redraw and reanalysis, he said.
Sheena said that Natera plans to launch the assay commercially before year's end, with a "select group of early customers."
Natera's vice president of marketing, Gautam Kollu, told CSN that the company has not decided on a price for the test yet, but that it will be "well within the range" of other commercially available non-invasive assays.
At last week's ASHG meeting, there was some discussion about the potential for non-invasive prenatal sequencing to diagnose many other disorders aside from aneuploidies (see story, this issue).
According to Natera's website, the company is currently conducting a retrospective trial evaluating its assay's ability to detect a broad group of chromosome abnormalities including microdeletions.