The Jelly Bean Conundrum

First an author's note: I am going to write my blog post for today very purposefully on the article I found most interesting, and NOT the two articles on laboratory culture which I though were needlessly pedantic, inauthentic to my experience in laboratories, and arrived at no useful conclusion. 

“The Molecularization of Race: Institutionalizing Human Difference in Pharmacogenetics Practice” by Duana Fullwiley is an article exploring what is often taken as a given in the current science of genetic disease, and that is accepting race as a valid qualifier of genetic data. You might have read before that there is no genetic basis for race, and accepted that as gospel, but then why do all this research into racially tailored pharmacology? What scientists mean when they say there is not genetic basis for race is that variation in human genomes between what we perceive as different races (Black, Native American, Hispanic etc.) is smaller than variation within that particular race. However, this does not mean the interracial variation does not exist. Thus we end up with these non-scientific categories defining and categorizing the data in these government-funded studies like the PMT Project Fullwiley discusses. This project originally had much less of vested interest in race, as Fullwiley describes, but then when they felt helpless to categorize the data, they reapplied and received funding for a new study where “racial purity” (term used by scientist involved with the study) of the subjects would be taken into account. And thus the PMT Project, whose goal was to study the frequency of certain SNPs within cellular transporter genes, ends up producing interpretations of data like that shown in Figure 2 of the article. This table uses race as the only social identifier to define the frequencies of the mutations in the gene, ignoring other potentially pertinent traits like age, or gender. And this study is not alone. Many similar studies for many types of genes and mutations are published, and many identify pertinent trends of genetic variation related to disease. Thus you have the justification for drugs like BiDil, an African-American specific heart medication. But can we really put race forward as a valid qualitative variable in these studies, even after it has been seen to hold little weight in our overall genome? Fullwiley doubts the power the social construct of race has over our DNA, and I too see potential for misleading results based on racial categories. The problem as I see it, is one of statistical significance. Most scientific studies utilize something called the P value, which represents the probability of your results occurring by chance with no underlying reason to evaluate the significance of their results. But if enough studies use the same potentially arbitrary categories, the P value itself can begin to be subject to the whims of chance. And that leaves us with, what I like to call, the Jelly Bean problem:

(image credits to xkcd.com, Randall Munroe)