Sequencing the world’s first vegan hamburger

tl;dr: we see DNA sequences from plants, yeast, and bacteria.

When you submitted your Seeq sample to be sequenced (not sequenced yet? Get the app here!), you probably noticed a question at the end of the instructions: “What did you eat and drink today?”

Why did we ask this question? We’re not trying to judge your junk food habits, I promise. Instead, we were motivated by an observation from some early experiments— specifically, saliva samples seemed to contain DNA from food particles, presumably from what people were eating prior to providing the sample. If this is true, then it might be possible to explore how your microbiome is influenced by your diet without ever asking you about your diet! That would be pretty neat.

Obviously I wanted to test this possibility on myself. So I enlisted a crack team of geneticists, enticing them with a free dinner:

Mmm, plant-based meats

We all headed to Momofuku Nishi, which we chose for one specific reason: it is the New York home of the Impossible Burger, claimed by its inventors to be “a delicious burger made entirely from plants for people who love meat”.

Indeed, all parties confirmed that the burger 1) was delicious and 2) tasted like meat. But was it really made from plants? After eating, in an egregious violation of the Seeq instructions and basic social norms, I ducked into the bathroom to spit in a tube, while one of our intrepid geneticists stuffed some burger into her purse for analysis back at the lab.

We then extracted DNA, threw it on a sequencer, and waited.

After we got my sequences from my saliva sample back [1], we removed all human and bacterial DNA sequences, and looked at what was left:

Far and away the most abundant DNA was from wheat, followed by tomato, potato, cucumber, and the yeast Komagataella [aka Pichia] pastoris [2]. I indeed enjoyed my Impossible Burger with the traditional tomato and pickle toppings and a side of fries, so the natural conclusion from these data is that the Impossible Burger itself is made from wheat and yeast. This seems to be in line with what is reported by the company (they don’t mention yeast, but perhaps this is how they’re growing the heme protein mentioned on their site [3]).

So what about the DNA sequences from the burger itself? Does the DNA profile of the food match the DNA profile inferred from my saliva?

For this analysis, we didn’t filter out any species since approximately all of the DNA should come from the burger:

Again near the top we see a lot of wheat and Komagataella pastoris DNA, and we can see traces of soybean as noted on the Impossible Burger ingredient list. Additionally, we see a couple of bacterial species [4] — Xanthomonas campestris, used in the production of the food additive xanthan gum, and a lactobacillus species, used in fermentation.

So what do I make of all of this? A couple key points:

1. DNA sequencing gives us some insight into the under-appreciated workhorses of industrial food production. We can see the genetic signature of the yeast that are (I presume) engineered to make key ingredients of the Impossible Burger, and the bacteria used in fermentation (exactly what Impossible Foods is fermenting isn’t obvious to me; historically fermentation is used for things like beer and kimchi rather than vegan meat) and the production of additives.
2. We can get at least some information about your diet from sequencing your saliva. This was the motivation for doing this analysis to begin with (ok, it was also an excuse to go to Momofuku Nishi), and now we know anecdotally that this seems to work! Using your data, we’ll take a closer look at this question shortly.

—

[1] Sequences describes in this post are all available here (file names should be self-explanatory). These are .bam files that can be manipulated with SAMtools. They are the output from alignment of sequencing reads to human genome build 37 with bwa mem.

[2] Note that the proportion of DNA sequences is not the proportion of ingredients/cells, since of course different species have different amounts of DNA. Wheat, for example, has a huge genome.

[3] Also they mention potatoes, but I unfortunately confounded analysis of the potato content of the burger in this analysis by eating fries as well. Maybe I should have resisted, but I’m only human after all.

[4] I confirmed that these bacterial species were also present in my saliva, along with a whole bunch of other microbes that presumably make up my normal mouth microbiome.