The data is available at http://citizenscience.microbiomeprescription.com/. I did some tutorials a few years back. These are linked to below:

• OpenSource Microbiome Project
• Uploading ncbi hierarchy data
• Computing Statistics
• Non Parametric Detection

For an example of checking any Artificial Intelligence Application, see Cross Validation of AI Suggestions for Nonalcoholic Fatty Liver Disease

Some challenges for the reader:

• Compute Percentile for each bacteria (taxon) in samples from the same lab
• Test the data if it is a normal distribution
• Run regressions between different taxon/bacteria using:
  • Raw Counts
  • Percentile
  • Which gives stronger results?
  • Reframe this using random forest and other ML techniques.

Then compare your results to that shown here for Clostridium butyricum (taxon 1492).

When you use percentile — what is the name of the resulting distribution? Why is that an advantage?

The model was built using Myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) for tuning. I recently did a Cross Validation of AI Suggestions for Nonalcoholic Fatty Liver Disease. Validation requires a lot of studies trying different things on a condition.

Next Project

Identify bacteria shifts for symptoms reported.

• Do for all labs first
• Filter by individual labs

Which approach gives better associations/models.