Sunday, February 18, 2007

Prenatal DNA testing: as simple as drawing blood?   posted by p-ter @ 2/18/2007 01:52:00 PM

Ever since the presence of fetal DNA in maternal plasma was demonstrated a decade ago, a number of teams have been searching for the best way to put this information into practice, so far with fairly disappointing results. The problem, of course, is that maternal DNA is also present in maternal plasma, and given that half a child's DNA comes from the mother, it's rather difficult to distinguish the two. The easiest applications, then, are when there is some DNA in the fetus that simply cannot be maternal--a Y chromosome, for example. Sex testing via this method should be rather trivial, though perhaps not that useful (the classic "look at the ultrasound and see if there's a penis" method also proves rather reliable).

After sex determination, testing for aneuploidies like Down's syndrome should be next lowest hanging fruit. Currently, the levels of a number of proteins are tested in expecting mothers, and those who have levels at the extremes of the distributions are advised to undergo amniocentesis to confirm or negate the results (the best tests now have a detection rate of about 80% and a 7% false positive rate, which, while not awful, certainly isn't diagnostic). As amniocentesis carries a non-negligible risk of causing a miscarriage, it would be desirable to get better results non-invasively.

Two new papers attack this problem from different angles--one uses paternal information to find SNPs that distingush maternal from fetal DNA, and the other cleverly uses levels of RNA from a gene expressed only in the placenta. Neither is going to replace amniocentesis just yet, but technology marches on...

Eventually, will it be possible to get an entire fetal genotype non-invasively? It's certainly possible, and this could have major implications for "neo-eugenic" practices. Currently, if a couple concieves a child naturally (no IVF), there are only a limited number of genetic tests it's worthwhile to do, as methods for obtaining the fetus's genetic information carry the potential cost of killing it. Once that cost is gone and our knowledge of genetics progresses a bit, parents could theoretically "screen" their children for whatever trait they desire without passing through the rather unromantic (and expensive) IVF step. Will many parents abort a child because they don't like its eye color? Doubtful, but some small percentage will choose to terminate pregnancies for reasons that seem rather cold-hearted.