Blood tests on an expectant mother, like the Natera Panorama, are now being used regularly to screen fetuses for chromosome abnormalities. At my wife's recent prenatal visit, she wasn't really even asked if she wanted the test; apparently it's standard in that office. I understand some of the science of how such a test could be performed. When the placenta implants, some of the fetal cells get into the mother's bloodstream, and even though the chromosomes from those cells are breaking down, these pieces can be scanned for the markers for genetic disorders and gender. But I don't understand how the fetus's maternal chromosomes can be identified against the background of the mother's genome.

For example, one of the things that they can rule out with pretty high certainty is Monosomy X: only one X chromosome and no Y. I can see that after determining the mother's genome, it could be determined if there was no evidence of an X or Y from the father. But how would they know if an X was not contributed by the mother. If they are looking through decomposing pieces of DNA in the mother's bloodstream, are they not guaranteed to find each of her X chromosomes, whether or not the fetus has a copy of either one? How can they determine that any piece of a maternal fetal chromosome really came from the fetus rather than from the mother?

Disclaimer: I do not endorse Natera over any of their competitors. I mention them by name only because they are the only ones I am familiar with.


As you correctly discerned, distinguishing the fetal cell free DNA from the mothers cell free DNA in the blood stream is one of the major challenges in this type of diagnostics.

Consequently, the results that can be directly linked to the paternal DNA are the most reliable.

In the case of the example you give, the high accuracy stems from the fact that in case of Monosomy X, there is no indication of neither a paternal Y nor a paternal X chromosome. While the Y chromosome is the most easily discerned, a paternal X chromosome could also be detected because in that case the cell free DNA would contain fragments of three distinct X chromosomes.

This publication summarizes the issue nicely:

Once the cell-free DNA fragments have been purified, small differences between the fetal and maternal DNA sequences are exploited in order to make a specific fetal diagnosis. To date, the majority of studies have focused on the detection of paternally inherited sequences that are entirely absent from the maternal genotype, such as those on the Y chromosome of male fetuses. This target is particularly attractive as it comprises a large portion of DNA which is not otherwise present in women.

Variable regions of repeated DNA (short tandem repeats or STRs) can be used to identify paternally inherited sequences, by comparing the number of repeats present with the paternal and maternal sequences. Paternal alleles on the autosomal chromosomes can also be detected, if they are known to be absent in the maternal genome, but this requires detailed sequence knowledge of the paternal genotype of interest, as well as detection methods that can distinguish DNA sequences that might only differ by a single nucleotide. Importantly, all these methods rely upon the fetus inheriting a uniquely paternal sequence that is conveniently located for a particular diagnosis.

The reliability of these tests can apparently be improved by enriching the fetal cell free DNA:

Selective enrichment of fetal DNA, based on a difference in the average physical length of fetal and maternal DNA fragments, which can be exploited to increase the relative amount of cffDNA. Fetally derived DNA fragments are generally smaller than those that are maternally derived, being predominantly ,313 base-pairs in length (Chan et al., 2004). Therefore, by using standard size fractionation to select only DNA fragments ,300 base-pairs, circulating cffDNA can be enriched such that it comprises around 70% of the total cellfree DNA (Li et al., 2004b), prior to detection and identification by either PCR (Li et al., 2005) or mass spectrometry (Li et al., 2006).

  • 1
    $\begingroup$ Tell me if I have this right: If we see no sign of a paternal X, then obviously we have monosomy X. But if we do see it, then in the enriched cellfree DNA, because an inherited maternal X chromosome would be represented throughout the entire sample of enriched cellfree DNA while a non-inherited maternal X would only be represented in about 30% of it, we should find approximately 3.3 times as much of one maternal X than the other to rule out maternal monosomy X. Or we would expect to see about the same amount of each maternal X if we have a problem. Is this about right? It's based on quantity? $\endgroup$ Feb 11 '16 at 16:31
  • $\begingroup$ I think in case of paternal chromosomes, the test is not only based on quantity but also on marker gene sequences (DNA fingerprints). From the blood sample, the DNA fingerprint of the maternal chromosomes are known from DNA contained in white blood cells. Now if the enriched cell free DNA contains DNA fingerprints of a X chromosome that does not belong to the mother, then its a girl, if we see DNA fingerprints of a Y chromosome, it's a boy and if we see neither, its a monosomy X. $\endgroup$
    – Thawn
    Feb 11 '16 at 19:32
  • $\begingroup$ However, excluding trisomy 21 has to rely to 50% on quantity of maternal chromosome 21 material because the trisomy could be caused by the child inheriting two chromosomes 21 from the mother which would only show up as an increase in chromosome 21 material in the cell free dna. $\endgroup$
    – Thawn
    Feb 11 '16 at 19:35
  • $\begingroup$ Oh. You're saying that monosomy X only happens paternally. It doesn't ever come from a missing X chromosome in the egg? $\endgroup$ Feb 11 '16 at 19:50
  • $\begingroup$ Honestly, I hadn't thought of that possibility. Of course, that could also happen... $\endgroup$
    – Thawn
    Feb 11 '16 at 22:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.