The patient has done another sample and this continues the analysis. The original post, A German CFS Patient Experience and Analysis from July 2021. I am always interested in seeing changes that occur from the AI suggestions, they worked for me –I am as interested in them not working (to motivate improving algorithms).

What has changed?

There are a set of automatic comparison tools between samples, so I will check those first

I went to compare taxonomy outliers for the two samples and found just two items were on both. There was a lot of changes. We can ignore blues with no data besides – they were all likely what I term transitory bacteria that comes and goes. What we are interested

The two items of focus are VERY high on both:

• (species) Paraprevotella xylaniphila
• (species) Porphyromonas canis

Other component comparisons:

• End Product — most were unchanged except for Polyhydroxyalkanoic acids which jumped from none to 98%ile
• KEGG Bacteria Products nothing in common, on first sample we had only low values at risk, on latest sample only high values at risk
• KEGG Enzymes — the same thing, earlier sample has all risks being low samples, latest high values
• KEGG Modules — earlier sample had no risk items, latest had one high value
• Medical Condition (PubMed) — nothing for either sample

The overall impression is that things have changed significantly. This does not say that the person is better — only that the suggestions have changed the microbiome

Non-Automated Comparisons

Next I will look at various aspects and do side by side comparisons. First, using Dr. Jason Hawrelak Criteria, shown below. The worst shift was for Blautia, the best shifts was for Roseburia and Faecalibacterium prausnitzii

				Latest	Sample	Earlier	Sample
Taxonomy	Rank	Low	High	Your Value	Status	Your Value	Status
Bacteroidia	class	0	35	38.369	Not Ideal	33.945	Ideal
Akkermansia	genus	1	3	0.21	Not Ideal	0.017	Not Ideal
Bacteroides	genus	0	20	25.986	Not Ideal	25.266	Not Ideal
Bifidobacterium	genus	2.5	5	0.03	Not Ideal	0.021	Not Ideal
Blautia	genus	5	10	11.557	Not Ideal	6.174	Ideal
Desulfovibrio	genus	0	0.25	0.062	Ideal	0.091	Ideal
Eubacterium	genus	0	15	0.008	Ideal	0.004	Ideal
Lactobacillus	genus	0.01	1	0.012	Ideal	0.017	Ideal
Methanobrevibacter	genus	0.0001	0.02	0	Not Ideal	0.052	Not Ideal
Roseburia	genus	5	10	5.166	Ideal	0.624	Not Ideal
Ruminococcus	genus	0	15	2.848	Ideal	3.792	Ideal
Proteobacteria	phylum	0	4	11.225	Not Ideal	5.753	Not Ideal
Bilophila wadsworthia	species	0	0.25	0.598	Not Ideal	0.848	Not Ideal
Escherichia coli	species	0	0.01	0	Ideal	0.022	Not Ideal
Faecalibacterium prausnitzii	species	10	15	10.612	Ideal	3.415	Not Ideal

Health Status changes were a toss up, but appears marginally better

• Latest: 0 Healthy, 4 unhealthy
• Earlier: 2 Healthy, 6 unhealthy

KEGG Suggested Probiotics

• Earlier Sample had 11 probiotics listed with a maximum weight of 12 (most of the time I see much higher weights indicating greater dysfunction)
• Later Sample had 0 probiotics – this is actually a rare occurrence which hints at a better balance microbiome

KEGG Suggested Supplements

• Earlier sample has 3 items suggested (using default percentile)
• Later sample has 0 items suggested (using default percentile)

Pub Med CFS Profile

• Earlier sample had 5 low matches, 5 high matches
• Later sample has just 4 high matches

Citizen Science Models

• Early Sample: Total Matching Bacteria :26. Very Strong: 15, Strong: 8, Weak: 2, Very Weak: 1
• Later Sample: Total Matching Bacteria :24. Very Strong: 19, Strong: 4, Weak: 1, Very Weak: 0
• Less matching bacteria (means some have disappeared), while strong/very strong are the same, the weak association have dropped

From looking at all of the available objective measures:

• The microbiome has changed (our primary goal)
• Many measurements show indicate that the microbiome is moving towards normal, none has a suggestion of things getting worst.

Remember our goal is not to attempt a one-step cure

Our earlier post used this diagram, we appear to have successfully moved along the path to recovery. We need to see what our next step is.

Symptoms Forecasts To Reported Symptoms

This reader was a good user and entered their symptoms – we have a 80% match rate. What is interesting is that one of the items not checked was a Keto Diet — this is interesting because a recent post dealt with a Keto diet resulting in Chronic Fatigue Like symptoms, with “Keto Flu” is some studies. A Keto diet is DEFINITELY NOT A SUGGESTION for this person, it will likely make things worst.

Updated Suggestions

For the earlier post, the consensus reports were not available. My intent is to run each of the matched predicted symptoms, the PubMed profile for ME/CFS and Citizen Science for ME/CFS – a total of 18 sets of suggestions will be merged. There are two ways of doing this:

• Only the Auto Checked (which picks a few best items)
• Auto Checked plus suggested one (marked with a light bulb).

This first pass is only with auto checked items (the most conservative approach) The bacteria selected are shown below, while there was some duplicates between symptoms, there was significant independence:

• Acidobacteria Too Low
• Anaerovibrio Too High
• Anaerovibrio lipolyticus Too High
• Bacteroides cellulosilyticus Too High
• Bacteroides finegoldii Too High
• Bacteroides salanitronis Too High
• Bifidobacteriaceae Too Low
• Bifidobacteriales Too Low
• Bilophila Too High
• Bilophila wadsworthia Too High
• Candidatus Phytoplasma Too High
• Clostridium malenominatum Too Low
• Deinococcaceae Too High
• Deinococcales Too High
• Deinococcus Too High
• Lelliottia Too Low
• Lelliottia amnigena Too Low
• Mitsuokella Too High
• Oscillospira Too High
• Parabacteroides johnsonii Too High
• Rhodocyclaceae Too Low
• Sphingobacterium bambusae Too High
• Sutterella stercoricanis Too Low

I noted that many are atypical bacteria that I do not see usually in pub med studies.

Safest Takes is full of items that I often have seen on my own (and other ME/CFS results).

On the Safer Take, we find a regular pattern of 3 takes and 1 oppose.

The avoid list is relatively short and most are from a single suggestion (which suggests that it may be ignored). high red meat is the greatest avoid, with saturated fats variation being right behind.

Done with Secondary Suggestions

Since I have just implemented this feature, I am curious about it’s impact. The number of bacteria almost doubled with a variety. A lot of bacteria appeared multiple time.

• Absiella Too Low
• Acidobacteria Too Low
• Acidobacteriaceae Too Low
• Acidobacteriales Too Low
• Acidobacteriia Too Low
• Adlercreutzia Too Low
• Adlercreutzia equolifaciens Too Low
• Anaerovibrio Too High
• Anaerovibrio Too High
• Anaerovibrio lipolyticus Too High
• Bacteroides cellulosilyticus Too High
• Bacteroides finegoldii Too High
• Bacteroides salanitronis Too High
• Bacteroides vulgatus Too High
• Bifidobacteriales Too Low
• Bifidobacterium Too Low
• Bilophila Too High
• Bilophila wadsworthia Too High
• Blautia hansenii Too High
• Butyricimonas Too High
• Candidatus Phytoplasma Too High
• Clostridium malenominatum Too Low
• Collinsella intestinalis Too Low
• Coriobacteriia Too Low
• Deinococcaceae Too High
• Deinococcales Too High
• Deinococcus Too High
• Eubacterium dolichum Too Low
• Eubacterium oxidoreducens Too Low
• Finegoldia magna Too Low
• Hymenobacter Too High
• Lelliottia Too Low
• Lelliottia amnigena Too Low
• Leptospira Too Low
• Leptospiraceae Too Low
• Leptospirales Too Low
• Megasphaera Too Low
• Mitsuokella Too High
• Oscillospira Too High
• Parabacteroides johnsonii Too High
• Peptococcaceae Too High
• Porphyromonas Too High
• Pseudanabaenaceae Too High
• Pseudonocardiaceae Too Low
• Rhodocyclaceae Too Low
• Rickettsiaceae Too Low
• Rickettsieae Too Low
• Selenomonas artemidis Too Low
• Slackia Too High
• Sphingobacterium bambusae Too High
• Streptococcus thermophilus Too High
• Sutterella stercoricanis Too Low
• Synechococcaceae Too High
• Synechococcales Too High
• Veillonellaceae Too High
• Veillonellales Too High

As expected, the suggests are similar

The risk items are similar with saturated fats dominating

Bottom Line

Earlier today, I updated this post Ketogenic diet did not help a health issue, it created one where the person ended up with a ME/CFS like scenario. Many of these suggestions were very similar. The main items suggested are:

• Fibre
  • Inulin like items (chicory, Jerusalem artichoke, inulin)
  • Fruit and legume fiber
  • Barley, Oats
• Cacao (usually I suggest working up to 85% cacao or higher chocolate bars)
• Apple
• clostridium butyricum (probiotics)
• rosmarinus officinalis (rosemary)
• B Vitamins (classic CFS MD recommended supplements)
• Vitamin D
• Soy
• Human milk oligosaccharides (prebiotic, Holigos, Stachyose)

In general, I favor “intact” fiber and herbs. I have seen several studies that the intact had a greater impact than the extract. Watch out for saturated fats — for example Pork is on the to avoid list

The excel file download (CSV) had only 291 items, less than other samples that I processed. I attached it as an example.

My usual advice is to take two weeks to sort out (and in some cases slowly move up to) the suggestions that are picked. It is helpful to check my Supplement Dosage page. That page is not numbers that I picked out of the air, but dosages used in various clinical trials (thus the dosages are likely safe, and more important, are sufficient to cause a change). Then at least 4 weeks for the microbiome to stabilize, then retest to find out the new status quo that will need to be adjusted.

For myself, I did notice a pendulum swing in my series of samples — an item was a take, then the next sample it was an avoid, then it became a take again and stayed as a take. The path is not always straight!