Grandma was right, eating your greens is good

Hey, do you like fruit and vegetables? Of course, you don’t, and if you do, there’s a high chance you’re recovering from a traumatic brain injury. For the rest of us, knowing what to eat is difficult. Everything that tastes good is linked to some form of cancer or crippling age-related disease. So, how do we digest the right information? The answer is, at least in part, using basic statistics.

Recently, as part of a 3-month rotation project. I was given the task of drawing links between what people ate and their general health. This included stuff found in our blood like sugar (glucose), what our bodies release to deal with too much sugar (insulin) and different types of cholesterol (good and bad). As for what people ate, I was looking at how many polyphenols they consumed. Polyphenols might sound like something that explodes on contact with water, but they’re actually a big family of chemicals found in plants. Funnily enough, most polyphenols in the UK come from tea, coffee and chocolate, which rather unsurprisingly reaffirms for me our position as a beacon of health on the world stage.

Looking at what foods people ate, and the polyphenol content of said foods, I was able to determine exactly how much of a specific polyphenol they consumed. Generally, consuming polyphenols was linked to lower levels of insulin in the blood. In particular, consuming luteolin was linked to the low insulin levels. Luteolin mainly comes from sweet peppers, which is good, because everyone loves a pepper. I mean, have you tried Mexican food? It crops up in broccoli, grapes, marrow, oranges, spinach and tea too. Another polyphenol called naringenin was linked to lower levels of fat in the blood and is mainly found in grapefruit, but most get it from oranges. I can only assume this is because grapefruits are relatively unwieldy compared to the humble orange. Try taking a grapefruit as a snack into work. You’ll get some bemused looks; some people awkwardly shuffling away from you and you probably won’t be invited to this year’s Christmas or non-denominational festive do.

Moreover, we tried to see if we could predict someone’s insulin level from their polyphenol consumption, which was quite a weak result. It indicated to me that a lot more dietary and health information is needed to predict such a thing. Still, it was a worthwhile thing to do because now when people ask me about research experience, I can say that I’ve used machine learning. This nearly always sounds exciting and impressive, despite the fact that I have no idea how it actually works. But such is the life of the lazy bioinformatician.

Finally, I know what you are thinking. I don’t need a scientist to tell me that peppers, grapefruit and oranges are good for me. Plus, I have an apple a few times a week, so I’m probably doing alright as far as my health is concerned. This isn’t accounting for my monthly trips to the gym either! Look, I’m not going to disagree. But you can’t deny it didn’t use up some of my time, and in the context of scientific research, isn’t that all we’re really striving for? I’m sure my granny would have agreed.