In this post, AI Computed Probiotics from Symptoms, we could calculate probiotics that could help for one symptom at a time for the general population. This is nice if you have just one symptoms and no microbiome details. See Ways of Choosing Probiotics for an overview on picking probiotics.

We can do better, a new page is up that will allow us to calculate the probiotics based on multiple symptoms PLUS your microbiome sample! In other words using all available information. (I will not create a page to handle multiple symptoms with no sample — you need to get a sample).

You must have entered symptoms for this to work. If not, you will see this appearing

After you enter symptoms, a page may appear like below

I should emphasis a few things:

• This is by retail probiotic name.
  • The probiotic must be available somewhere in the world. It may not be available where you live
  • If you wish to know which species are in the probiotic, just click the name.
    • Probiotics with the same numbers are likely the same species (i.e. no difference)
• The Weight is an estimate of how much of the missing enzyme it will provide (weight is based on odds)
• Enzyme Means is the number of Enzymes that will be provided by it
• Species is the number of different species in it.

Practical Example

Using the above example, the person founds that Prescript-Assist®/SBO Probiotic is either not available or too expensive (watch costs!) and proceeded down the list:

• miyarisan (jp) / miyarisan is available by a Japanese website at a reasonable price.
  • Record the species, in this case: clostridium butyricum miyairi
• The next 4 are all the same species, lactobacillus plantarum, the best buy for this person was CustomProbiotics.com / L. Plantarum Probiotic Powder (remember prices vary greatly from place to place)
• Next, they work down the list to find something that contains neither of the above (different strains, different enzymes)… the winning bacteria is: lactobacillus salivarius
• Moving on, we find ImmuneBiotech Medical Sweden AB / GutMagnific® contains 2 of the above, so we skip it… the next one is lactobacillus rhamnosus (37.8) found in 8 different products.

Cross Validation

We have other ways of suggesting probiotics

• Looking them up by research from this page [Search for probiotics by studies] – unfortunately, this will not giving a ranking
• Using the KEGG based calculation without using symptoms:
  

• The third way is by suggestions — here, the choice of bacteria selection can result in a wide variation of probiotics suggested and contradictory results as shown below:

What to do with contrary results?

We potentially have 4 opinions

• From Symptoms + KEGG
• From KEGG alone
• From the literature
• From suggestions

To take or not to take should be done on consensus (i.e. ideally 3 says to take). Of the above methods, the one with the weakest quality of data is from suggestions (because it is so dependent on studies being done! ). For the one in conflict, lactobacillus salivarius (AKA Ligilactobacillus salivarius), there were studies found in the above link (strain specific for retail probiotics), NOTE: I missed them on the first pass because I did not enter the name in “Search for” and had left the default ‘constipation’ there

One of the symptoms was brain fog and depression. “sad mood” is a sufficient match.

To translate the methods into human “detective” terms

• KEGG — DNA is a match
• KEGG + microbiome Sample — DNA and video is a match
• Researched Studies — Profiling by race, sex, age etc is a match (studies are done on populations, not individuals) – it is truly “bacteria profiling”
• Suggestions based on bacteria picked — close to setting up police stops to detect drunk drivers. The number of people arrested depends on time, location etc of where the stops are done.