Is there a gene for oversimplistic analysis?

Earlier today I had the privilege of attending* the annual Sluckin Memorial Lecture given by eminent Oxford neuroscientist and academic blogger Professor Dorothy Bishop. Dorothy’s theme was ‘Developmental dyslexia and other neurodevelopmental disorders: Distinct syndromes or part of normal variation?‘. There was much in the talk worthy of blogging here, but since I’ve got a stack of final year dissertations to mark I will, for the moment, limit myself to reflections on one point that she raised.

Slide 19 in a presentation by Dorothy Bishop available on Slideshare

Slide 19 in a presentation by Dorothy Bishop available on Slideshare (click image for link)

As with many conditions in the genomic era, there is a desire to find the underlying genetic ’cause’ for dyslexia. This search is not without justification. For example, classic comparison of monozygotic twins (“identical” twins, i.e. same genetics, notwithstanding any epigenetic influences) and dizygotic twins (“non-identical twins”, no more genetically related than any brother or sister) strongly implies that there is a genetic component to dyslexia.

There is stronger evidence than this, particularly for a correlation between dyslexia and the catchily name gene DCDC2. A 2005 paper in the Proceedings of the National Academy of Science, a “Premier League” academic journal, showed a link between specific mutations in this gene and reading disability. A subsequent paper by Tom Scerri and colleagues (including Dorothy) found that a particular Single Nucleotide Polymorphism (a SNP, i.e. a particular base change difference in the DCDC2 gene) was associated with 31% of dyslexics. It was also found in 23% of the control (i.e. non-dyslexic) group, but nevertheless the difference the two is statistically significant (p = 0.005).

So there is some good scientific basis to the notion of a genetic factor in dyslexia. The problem comes, however, with the suggestion (as in this press release) that there might be mileage in using the existence of this link as an early warning test for dyslexia. Dorothy elegantly used the DCDC2 story as a demonstration of the dangers of an oversimplistic application of genetic association data. [Slides from a very similar presentation by Dorothy (including the relevant information on slides 14-19 , can be found here.)]

Suppose (not unreasonably) that 10% of the population are dyslexic. If 31% of them have the particular SNP associated with dyslexia then that would equate to roughly 3 people in a population of 100. Only 23% of the non-dyslexic population have the same mutation which, we have seen, is a statistically significant difference. However, 23% of 90% of the population equates to roughly 21 people in our population of 100. So, even though the gene is associated with dyslexia, a genetic screen based on this fact will throw up 21 false positives for every 3 real dyslexics identified (and the majority of dyslexics would not be picked up by the test). The science, good science, has therefore been stretched beyond its legitimate uses.

I was reminded of the condition “Brain Overclaim Syndrome” coined by Stephen Morse of the University of Pennsylvania. Morse was commenting on the related tendency, particularly in the context of criminal responsibility, to take information from brain studies and to apply it too simplistically to what is, in reality, a far more complex situation.

We live in truly exciting times regarding discoveries in both the genetics and neurophysiology of brain-related conditions. Nevertheless we must all be cautious so that this knowledge is not abused in the rush for simple answers to sophisticated questions.

(* In fact I was oblivious to the lecture being on, but I came down the stairs in my building at work and saw on the big screen in the overflow area that the lecture had just started. As someone who follows Dorothy on Twitter I knew it would be an excellent talk so I blagged my way in. Definitely a more profitable way to spend an hour than whatever I might have been intending to do).

Advertisements

4 Comments

  1. V glad that the Sluckin Lecture gave us the opportunity to get acquainted, and delighted that you enjoyed the lecture.
    The overselling of genetic findings is a bit of a hobbyhorse of mine, and I have various blogposts about this, including these two that may be of interest to your readers:
    http://deevybee.blogspot.co.uk/2013/01/genetic-variation-and-neuroimaging-some.html
    http://deevybee.blogspot.co.uk/2012/04/getting-genetic-effect-sizes-in.html

  2. […] genetics stories: Is there a gene for oversimplistic analysis? and Everyone is related 1000 years […]

  3. […] You may also be interested in the post Is there a gene for oversimplistic analysis? […]


Comments RSS TrackBack Identifier URI

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s