In this challenge we’re diving deep into the microscopic interactions between you and the trillions of microbes that call your body home.
In the MPEG-G Microbiome Classification Challenge, you helped classify sample types—but let’s push the boundaries.
➡️ What else is worth classifying?
➡️ Could you distinguish between IS and IR samples?
➡️ What might that tell us about disease, treatment, or even aging?
If you could take one swab from anywhere on your body—what could it reveal about your health, your immune system, your lifestyle?
These aren't just theoretical questions. This challenge is an opportunity to build something that could contribute to real-world discoveries in personalized medicine and immunology.
👀 Pro tip: Read the accompanying paper. There’s a wealth of insights in there that could help shape your approach—and spark ideas for classification tasks beyond just sample type.
It's a hard read but I did it so I believe in you!
Hi Amy,
Thank you for your efforts. I have bit of an issue with Tracks 3 and 4, which seem to indicate that the metagenomic samples and its accompanying cytokine data are time varying. However, looking through the dataset, I cannot find any timestamps or something to indicate its time-varying nature.
Samples that have time stamps associated with them are in the cytokine data frame in the last column.
Note that when a sample is taken, up to 4 samples from the four different locations are taken at the same time AND a sample of blood is taken. The cytokine data is pulled from the blood data.
I hope this helps.