Feeding the genome

Photo by Thinkstock.

Photo by Thinkstock.

When the Human Genome Project wrapped in 2003, we assumed this ginor­mous data set would pro­vide the much needed parts-​​list to fill in the blanks of human health and dis­ease. But in the last 10 years it’s become exceed­ingly clear that things are just not that simple. Yes, our genes are obvi­ously more than a little impor­tant in deter­mining who we are, but a lot of other fac­tors are, too. You prob­ably wouldn’t be sur­prised to hear that what we eat is a big one.

Health sci­ences pro­fessor Katherine Tucker has ded­i­cated her life to under­standing how nutri­tion affects us and, in par­tic­ular, how our nutri­tion can be cul­tur­ally deter­mined based on our diets. Although indi­rectly, our genetics are con­nected to our cul­ture as well. Now don’t get too worked up, I’m not saying there’s a gene for faith or any­thing like that. But there are plenty of exam­ples of cer­tain gene muta­tions being more preva­lent among some pop­u­la­tions than others. For example, a muta­tion to the BRCA1 gene, which is asso­ci­ated with increased risk for breast and ovarian can­cers, is more common among white women than Asian women. A muta­tion to the ABCA7 gene seems to be more often asso­ci­ated with Alzheimer’s dis­ease among older African-​​Americans than whites.

Tucker is part of a growing field called nutrige­nomics, in which researchers are studying how the nutri­ents we con­sume affect gene expres­sion. What does that mean exactly? Here’s one example: There are three pos­sible vari­ants of a gene called APOE, one is asso­ci­ated with higher levels of low-​​density lipopro­tein cho­les­terol (LDL-​​C, the “bad” kind), one is asso­ci­ated with mod­erate levels of LDL-​​C, and one is asso­ci­ated with lower levels. But throw a bunch of high-​​fat food into the mix and the inter­me­diate variant becomes a bigger pre­dictor of high LDL-​​C. So, just because you have a gene asso­ci­ated with some phys­ical char­ac­ter­istic, what you eat may change things completely.

It seems if we can nail down the sequence of the 3 bil­lion nucleotides that make up the human genome, we should have no problem dealing with a little Vit­amin A. But it turns out this is a ridicu­lously com­plex ques­tion to study. There’s a number of rea­sons for that, and one, as you may be guessing, is the great vari­ability in the way we eat.

Say you want to figure out how zinc intake affects gene expres­sion. You might first assume that the real hairi­ness of this chal­lenge would be looking at zinc’s inter­ac­tion with all 20K genes, but due to some serious tech­nical advances, this is actu­ally the “easy” part.

Because of the speci­ficity of most gene x nutrient inter­ac­tions,” write Tucker and her col­leagues in a recently pub­lished paper, “valid data are needed for nutrient intakes at the indi­vidual level.”

This basi­cally means that the genetic effects of your zinc intake might not look the same as those of my zinc intake. In order to get mean­ingful infor­ma­tion about how an individual’s diet affects his or her risk for dif­ferent dis­eases, for example, you need to look at many, many, many.…many dif­ferent individuals.

Okay, sure fine. If we can look at 20K genes then 20K people should be easy enough. True…and, well, not true. You can ask as many people as you want how much rice they eat, knowing that the whole grain kind is high in zinc. But rice means dif­ferent things in dif­ferent cultures.

To some, rice means sticky white rice cooked with a little salt. Others eat a ton of butter or oil with their rice, so now you have to think about how fat will interact as well.

There are three main ways nutri­tional epi­demi­ol­o­gists like Tucker ask people about their diets. They can ask what you ate yes­terday, but that’s not usu­ally a good indi­cator because you may have eaten a huge slice of cake yes­terday, some­thing you only do once a year on your birthday. They can ask people to keep a record for a week or more, but this is time con­suming so people don’t love doing it, not to men­tion the fact that keeping a diet journal is known to affect the way one eats. The third way is to admin­ister a “food fre­quency ques­tion­naire,” wherein you answer a whole bunch of ques­tions about how often you eat dif­ferent foods. FFQs, as they’re called, have to be specif­i­cally tai­lored to dif­ferent cul­tures in order to get accu­rate nutri­tional data, again: rice means dif­ferent things in dif­ferent cul­tures. So if you’re com­paring a bunch of FFQ results across pop­u­la­tions, the cor­re­la­tions with genetic fac­tors and out­comes are prob­ably very skewed.

So, as Tucker and her co-​​authors repeat­edly point out, the need for better, more effec­tive dietary assess­ment methods is crit­i­cally needed before anyone can reli­ably con­sider how nutri­tion and genetics are con­nected. We know that they are, of course, but to do any­thing real with that knowl­edge there needs to be a heck of a lot more stan­dard­iza­tion among the pro­to­cols, which will require some cre­ative thinking on the part of the people per­forming these studies.

Tucker has already begun col­lab­o­rating with researchers in Northeastern’s Per­sonal Health Infor­matics pro­gram to think about using tech­nology to stream­line the process. As becomes clearer every day, crowd­sourcing data through the Internet and our smart­phones is an incred­ibly effi­cient way of col­lecting data. You just have to ask the right questions.