A reader on Facebook requested this data (since I am likely the only one that has the data that can speak of it). Here’s the charts – Have fun interpreting
A reader message me about Kefir. My usual response is “You do not know what you are getting”. While for a person with near normal health, it likely does some good (keeping with the concept of hygiene hypothesis), this is not so clear for more severe dysbiosis of the gut.
“Kefir grains consist of complex symbiotic mixtures of bacteria and yeasts, and are reported to impart numerous health-boosting properties to milk and water kefir beverages. ” [2022]
Which bacteria could be in Kefir
Typical Kefir Label – What are these cultures?
There is no legal requirement to report the name of the bacteria in the kefir, nor the genus, species or strains. Each batch may have a significantly different mixture of bacteria. You want live and active cultures? Go to the forest and eat a spoonful of dirt!!
Kefir is not a precise product, even in the vaguest terms. I am bias to juggling as few balls at a time as possible… Kefir feels like working with all of the balls in an IKEA kids ball room. It’s spinning the barrel of a bacteria roulette — especially since the same product from the same manufacturer may be different in the next batch. For grown at home kefir, the variability will be a lot higher.
I am not a licensed medical professional and there are strict laws where I live about “appearing to practice medicine”. I am safe when it is “academic models” and I keep to the language of science, especially statistics. I am not safe when the explanations have possible overtones of advising a patient instead of presenting data to be evaluated by a medical professional before implementing.
I cannot tell people what they should take or not take. I can inform people items that have better odds of improving their microbiome as a results on numeric calculations. I am a trained experienced statistician with appropriate degrees and professional memberships. All suggestions should be reviewed by your medical professional before starting.
I have done everything as planned since your first review
I maybe went up from 15% to 20%. I was able to reintroduce some new activities, but still am lying in bed most of the time. Also taking piracetam seems to help.
I still won’t be able to do the analysis myself.
COMMENT:ME/CFS patients are a priority for me because I personally understand their brain fog and cognitive impairments from past experiences.
I have had COVID in the meantime in case that matters for your analysis, but I did not notice any changes afterwards.
Analysis
Given the recent post for another ME/CFS person who had COVID too, with the result that their microbiome became a good match for long COVID and a poor match for ME/CFS, this was my first question. Fortunately, the sample was done via Biomesight, he did not needed with FASTQ files and transferring them. To keep the story short, I looked at his shifts compared to annotated sampled and compare to literature from the US National Library of Medicine nothing shifted between the samples. There is no shift towards Long COVID from ME/CFS in this case.
Comparing Samples
I do not know the answers. I have a model. Models often need adjustments so comparing samples (for better or worst) in a consistent manner is part of my learning process.
First thing we see a dramatic change with rare bacteria being seen much more often and common bacteria less often. There are more genus seen (184 vs 141) and more Species (230 vs 161) but this may be due the better sample reads in the latest sample (82,102 reads versus 55,117 reads).
Percentile
Latest Genus
Latest Species
Earlier Genus
Earlier Species
0 – 9
47
65
2
4
10 – 19
23
27
10
16
20 – 29
19
16
16
11
30 – 39
13
17
12
13
40 – 49
15
18
13
14
50 – 59
16
17
23
32
60 – 69
15
22
14
19
70 – 79
17
22
13
14
80 – 89
13
16
24
22
90 – 99
6
10
14
16
Average
18.4
23.0
14.1
16.1
Std Dev
11.0
15.4
6.2
7.4
Hawrelak’s criteria was 95.6%ile for both samples.
Potential Medical Condition dropped from 7 to 1. With Obesity being in common.
The person feeling subjectively better and doing more activities
Most of the other measures are the same or difficult to interpret. There is one possible concern, the high levels of Prevotella copri is an indicator of mycotoxin, typically from moulds and fungi. Considering that the time between the samples was winter with close windows and heating — there could be an environment issues here – so lots of fresh air may be good.
Over to Suggestions
There are various algorithms to suggesting probiotics, the strongest results are for:
I ran a few ways of picking bacteria based on Bacteria (and not genes) and lactobacillus casei kept was the top in the consensus report (overall and in terms of probiotics)
Unfortunately, some of the items have no studies. Given that the suggestions are based solely on bacteria with no knowledge of the diagnosis, the convergence with the literature suggests that the suggestions are very appropriate. Two different roads came to the same conclusion. In data science this is sometimes called “cross validation”. In Scotland, “O ye’ll tak’ the high road, and I’ll tak’ the low road, And I’ll be in Scotland a’fore ye,”
I looked at the antibiotic list for the latest sample and the top two are typically used for ME/CFS:
And interesting that several others often used are NOT recommended: azithromycin (which is a macrolide ?!?), minocycline [2021], fluoroquinolone, doxycycline.
ME/CFS is a heterogeneous condition with a wide variety of microbiome dysfunctions. I believe that using the microbiome to target the best candidate antibiotics is the rational way to proceed.
Question: Sadly I do not tolerate chocolate, but I will try it out again.
Answer: These are suggestions, do only what you are comfortable with. Nothing is required. The chocolate issue is interesting, my daughter does not tolerate most chocolates, she discovered that it was the type of sugar (i.e. made with liquid sugar / liquid glucose — adverse reaction) made with solid sugar — happiness. See Health effects of glucose syrup
If you try again, you may wish to determine the type of sugar actually being used first.
Question: Is there no avoid list?
Answer: Yes, in the download, any item with a NEGATIVE value in the priority is an avoid
Avoid Cnt is the number of time that suggestions placed it into avoid
Question: Is 1 capsule of Equilibrium per day really enough?
Answer: I honestly do not know. There is no literature to work from. If you take more, than separate them (i.e. 12 hrs apart)
Question: It seemed whenever I took turmeric that I was getting more nervous and anxious. Still take it now and then?
Answer: As above, do only what you are comfortable with — there are hundreds of items listed. Anxiety is contrary to the effects of turmeric / curcumin reported in the literature [2021] [2019] [2018] [2017]. If turmeric is causing die-off of bacteria that causes vascular constriction, that would result in anxiety. If you tolerate aspirin or niacin (flushing type), then try taking those with the turmeric.
I am not a licensed medical professional and there are strict laws where I live about “appearing to practice medicine”. I am safe when it is “academic models” and I keep to the language of science, especially statistics. I am not safe when the explanations have possible overtones of advising a patient instead of presenting data to be evaluated by a medical professional before implementing.
I cannot tell people what they should take or not take. I can inform people items that have better odds of improving their microbiome as a results on numeric calculations. I am a trained experienced statistician with appropriate degrees and professional memberships.
Backstory of Latest Sample
In light of your recent few blog posts about uploads without many microbiome shifts to work with, I was thinking this could be a beneficial walkthrough video for what seems to be the opposite.
I was doing pretty well on my antibiotic rotations (mainly tetracycline two weeks on, two weeks off since Aug of 2021) until Feb or so when I had a major crash / flare that I’m still suffering from.
I did have a very mild case of Covid in mid January that felt no worse than a regular cold.
But from what little I can parse from this sample, it seems I may be struggling with long Covid. I say little, because my brain fog is extremely dense.
And all of the results I’m getting for this sample via your site seem so drastically different from what has been going on over the last 7 years (my oldest sample is from 2015).
Comparison of samples
This person has samples going back to 2015 using uBiome. Unfortunately for comparison we need to keep to the same lab (why? read The taxonomy nightmare before Christmas…).
Jason Hawrelak Criteria etc
We finally see an improvement with Jason’s criteria. We also may be seeing more diversity with the increase of Genus and Species found. I say may because this could be a side-effect of a low raw count in some samples.
Date
Percentile
Unhealthy Bacteria
Genus
Species
2022-04-11
98.8 %ile
8
220
303
2022-01-11
89 %ile
11
89
141
2021-03-09
89 %ile
8
108
153
2020-05-27
89% ile
7
153
223
Finally, we have a significant improvement
Expected values ar 10% for each line
I decided to look at the raw reads (which are captured from Thryve and Biomesights)
Sample Date
Raw Reads
5/27/2020
43311
3/9/2021
29247
1/11/2022
17630
4/11/2022
153194
The cause of the jumps above may be the number of reads from the sample
This lead me to look at what typical raw counts are from Ombre/Thryve
To find the raw counts for your sample, open the csv and look for this line
What is the consequences? It means that rarer bacteria may be ghost-like, appearing or disappearing from sample to sample. This adds let one more layer of fuzziness to doing analysis and generating suggestions.
First Question: ME/CFS or Long COVID microbiome or both?
This person uploaded the Ombre FASTQ files to BiomeSight so I may used data from the Long COVID study there. Both condition present similarly, I am curious to see if we have sufficient reference data to decide which condition is a better match.
The table above hints that he is at present much closer to Long COVID than ME/CFS.
I am not sure about the political correctness of saying “Congrads! You no longer have ME/CFS, you have Long COVID!” is what the microbiome reads like.
What is interesting is that the microbiome constantly shifts/evolves, with Long COVID the infection is constant and the duration since the infection is short — hence less evolution of the microbiome over all patients. With ME/CFS the triggering infection possibilities are huge with 20, 30, 40 years of evolution of the microbiome — hence patterns are diffused by time and original infection.
Looking at deficiency of compounds produced, we see a dramatic drop from the previous sample suggesting that bacteria are getting the needed inputs for correct functioning.
Sample Date
1%ile
5%ile
10%ile
5/27/2020
4
14
60
3/9/2021
2
14
16
1/11/2022
197
233
244
4/11/2022
6
28
52
Kegg Compounds below %ile shown
Where do we go from here
I am going to do consensus, but do only 3 items:
Hand Picked Bacteria using the study in progress data using BiomeSight (16 bacteria)
Using US National Library of medicine filter to Long COVID using BiomeSight and Box-Whiskers (14 bacteria)
Using US National Library of medicine filter to Long COVID using Ombre and Box-Whiskers (14 bacteria)
The consensus is below as a download. Since antibiotics are being prescribed at present, I included that in the suggestions criteria.
Why did I focus on the ME/CFS ones? Path of least resistance for the prescribing MD – the MD accepts ME/CFS and thus will have low resistance to prescriptions often used for ME/CFS. Asking for them for Long COVID could get rolling of eyes…. As always, we are using these off-label for their computed microbiome effect. For the prescription items, I would suggest rotation (one item for 10 days, then a 0-10 day break, then another item (or repeat if limited to one item).
This post is intended for researchers by pointing to bacteria whose genetics are likely significant for long COVID. The raw data is below. Preliminary z-scores indicated that they are significant (Pr < 0.01) and no filtering has occurred for False Detection Rate. Users are advised to perform their own statistics.
Note: These results are lab-specific, using the data provided by BiomeSight.
Some recent work has identified bacteria that are associated with Autism. For a summary of method, see this post. The following are the list of bacteria seen with Ombre/Thryve samples that are annotated with Autism. There are not sufficient samples yet for specific autism characteristics – so please check your uploaded samples and update the symptoms.
These are bacteria that you want to reduce (with one caveat — the suggestions algorithm requires the percentile to be 50%ile or more). How to hand pick them? See below the list.
Note: you may only have a few of these. They are shown in the same sequence as seen on Microbiome Tree. The LAST item is what was found to be statistically significant.
This is a technical note. Recently I came across this doing analysis of Long COVID data.
Thermosediminibacterales(order)
With Long COVID: 55/152 samples or 36.2%
Reference (excluding Long COVID samples): 72/996 or 7.2%
This present an interesting insight on possible blinkered thinking when seeing such data. Some examples are:
Don’t brother looking —
It’s a rare bacteria (just 7% of people have it…)
It does not occur in most Long COVID patients, not interesting
I computed the means and standard deviations, and the difference is not sufficiently significant, so do not mention
My take is simple, it occurs FIVE times more often. I view microbiome dysbiosis are the result of the “perfect storm” or should I say “imperfect storm”. The wrong concentrations of compounds and enzymes coming together from a host of bacteria. With that dysbiosis view, a rare bacteria oddity like this, hints at a subset. This is contrary to the common view that dysbiosis is caused by a single or small group of bacteria and you can make simple either/or decisions based on their presence or lack of presence.
In the case of the long COVID data, I observed some odd (by traditional thinking) situation. A few examples:
A 10 fold difference of frequency with the higher frequency having a higher average – the traditional expectation. More of this bacteria is growing, hence we find more often.
A 10 fold difference of frequency with the higher frequency having a lower average, with statistical significance. This is what stopped me to re-examine my perspective, including the need to re-evaluate some blinkers.
The natural question: Determining Significance!
For most people dealing with biological data, presence or non-presence is typically a dependent factor. For example, here are some means for bacteria with the outcome being Crohn’s disease detected or not (the control case). The data will often be dropped into logistic regression.
I went back to flipping bias coins thinking and raise a beer to the memory of Bernoulli. In the above case, the expected bias is that 7.2% of the time the coin will land with a head. We try a new coin and toss it 152 times and get heads 36.2% of the time…
The hypothesis to test is whether the coin is equivalent?
The standard deviation of the population is a simple calculation – except we need to change .50 to .362 in formula below. (P.S. The Std Dev of the population is about 1%, so a range of .342 to .382 could be tried safety)
From http://20bits.com/article/hypothesis-testing-the-basics
The result is a z-score of -7.43, or clearly significant well beyond a 0.01 level. Thus the presence or lack of presence is statistically significant and should be included in any analysis (but rarely seems to be in most papers)
One of the goal of Microbiome Prescription is to stay true to source data / study. There are many studies that deal with a diet style or atypical food elements, like ‘high milk fat’. Below these wide sweeping terms may be concrete specific items that are reported in a different manner. A simple examples:
Take Vitamin B2 (Riboflavin). Milk and beef are significant contributors
Underneath the covers of this complex microbiome engine in the human body, the impact of more beef or more milk is an increased availability of Vitamin B2.
Diets are complex concepts subject to regional interpretation. A high beef diet means more beef than a typical person… so how much is that [source]?
If you are in China, it’s more than 1 pound of beef a month.
If you are in Russia, it’s more than 2 pound of beef a month.
If you are in USA, it’s more than 3.5 ounces of beef a day (so, more than a MacDonald’s Quarter Pounder every day).
If you are in Uruguay or Argentina, it’s more than 5.5 ounces of beef a day.
When we go over to items like a Mediterranean Diet, often it can mean many things with a wide range of contents. Both of the following would meet that criteria for many people:
One serving of cereal, two servings of citrus fruits, one servings each of eggplant, okra , green beans
13 servings of cereal and breads, one half apple, five servings of potatoes, 3 servings of carrots, 1 serving of onions.
The MedDiet contained three to nine serves of vegetables, half to two serves of fruit, one to 13 serves of cereals and up to eight serves of olive oil daily. It contained approximately 9300 kJ, 37% as total fat, 18% as monounsaturated and 9% as saturated, and 33 g of fibre per day.
The majority of studies emphasized the same key dietary components and principles: an increased intake of vegetables, wholegrains, and the preferential consumption of white meat in substitute of red and processed meat and abundant use of olive oil. However, the reporting of specific dietary recommendations for fruit, legumes, nuts, bread, red wine, and fermentable dairy products were less consistent or not reported
To me, a medDiet is eating traditional Greek — stuffed grape leaves, Tomato Fritters, etc with a glass of Ouzo [example] – in my younger days while I was teaching, I would have this 3-4 nights of the week.
At this point, we find that most studies involving diet deteriorates into vague hand-waving.
Can you use diet style?
This is a two sided coin. If you take recommendation for items like Luteolin, it can be translated into diet such as more celery seed, olives, blueberries. Quercetin into Cranberries and Blueberries. etc. While a high meat diet is vague — does it mean beef? pork? chicken? fish? – how much?
A logical solution is to decompose the diet into an itemized list of what the diet means by component. Then using the wonderful databases at the US Department of Agriculture develop a profile of what you are getting with this style of diet. Usually there are multiple diet suggestions, so you need to intersect them to get the true bottom line on what the diet changes should be.
Bottom Line — Use Diet Style with caution!
IMHO, it is so close to saying “Buy tech stocks for your retirement”. Without doing due diligence, you may end up with a worthless portfolio. At the bottom of the suggestions is a Flavonoid section which could be translated into food specific items.
This is an interesting case which appears to illustrate well that microbiome-agnostic prescription of antibiotics can produce horrible results. Doing a yearly 16s microbiome test will allow you to potentially negotiate with your MD to pick antibiotics that both address the MD concerns and potentially improve your microbiome as a side effect. See this post: Negotiations with your Medical Professional
My backstory:
I have used FQ antibiotics many times in the last 15 years for Chronic bacterial prostatitis..
During the last few years I was diagnosed with diverticula and had an episode of diverticulitis 3 years ago which also required antibiotics.. In the last 2 years my bloating was so severe that I was like a pregnant woman.. I am a male 40 years old.. So last July I went to the beach and caught E.Coli once again from the water or the beach.. This gave me acute infection with fever the next day.. This is where the drama starts as I ended up going to 4 different labs giving me different results and switching antibiotics for 5 months.. My gut was so bad that I’ve spend one night at the WC and another day I was stuck in traffic and I didn’t come back in time.. So embarrassing..
So January I stopped the FQs since I got a severe reaction with a set of symptoms that almost took my life.. My calf tore while being in bed, not even walking, swollen joints with fluid, tinnitus, diarrhea for 1 month, stomach ache and spasms, neuropathy, brain fog, insomnia and more..
I was sure that everything started from my gut, something triggered auto-immune along with toxicity from the drugs.. 2 months in bed.. 4 months and I barely walk with many symptoms.. What saved me initially I think was homemade Kefir I had and making myself..
Then I did the test at Biomesight and understood why and what happened.. Now I know very well that life or death starts from the gut..
Current State
First, I like to get a feel for where the microbiome is at from a high level. Looking at the usual health measures:
Dr. Jason Hawrelak Recommendations guidance puts the person at the 35%ile, definitely in the concerning space
On the Potential Medical Conditions Detected, 14 items were flagged, again concerning
In the Bacteria deemed Unhealthy list, the following stood out
Looking at the distribution by frequency, nothing really stands out.
Percentile
Genus
Species
0 – 9
14
18
10 – 19
19
34
20 – 29
19
19
30 – 39
12
14
40 – 49
15
7
50 – 59
9
15
60 – 69
13
13
70 – 79
8
15
80 – 89
10
9
90 – 99
15
18
Looking at the antibiotics list taken, I went over to the Antibiotics List for MDs page for this sample. We are using this to see which antibiotics likely helped the dysbiosis of the gut to happen.
The following were the antibiotics that he had been prescribed. I put after each the positive and negative estimates from the above page. We see a -.266 for something taken for 84 days…
In this case, it is clear from the data above that the antibiotics were a factor for his problems. if he must take antibiotics again (or can persuade his MD to do a trial), the best ones suggested by the Artificial Intelligence algorithms are:
rifaximin (antibiotic)s (1)
metronidazole (antibiotic)s (0.887)
ampicillin trihydrate (antibiotic) (0.834)
Action Plan Going Forward
The KEGG AI Computed Probiotics had the HIGHEST VALUES that I have ever seen with the top items being, I would go for three of these (2 weeks of one, then rotate to the next, repeat): Something that lists bacillus subtilis as the first ingredient, miyarisan (jp) / miyarisan, something that is just lactobacillus plantarum (i.e. 299v)
For supplements, we have (even at 20%) a short list. Usually supplements can be taken consistently.
beta-alanine – Percentile: 5.2
Glycine – Percentile: 3
L-Cysteine – Percentile: 10.4
L-glutamine – Percentile: 15.5
Magnesium – Percentile: 3.7
Molybdenum – Percentile: 0.9
Building Consensus Suggestions
Remember, no one knows how to pick the best bacteria to target. We apply multiple criteria and then work from what is agreed upon with the different approaches (i.e. consensus).
Use JasonH (15 Criteria) – 11 bacteria picked (and the same for the other ones at the top of this list)
from above: magnesium (found in consensus too with positive value), molybdenum and from consensus: selenium and zinc. But no iron supplements.
Probiotics
All of the items cited above are on the to take list (and many more), but the ones above are IMHO, “double blessed”
A reminder, the items are based on the term that various studies used. In some cases, there can appear to be contradictions. In some cases this could be due to what was measure or not measured in the study as well as sample size. We do not know what is “right” when this happens, it drops into a state called “indeterminate”. There are some of those here, but also some very clear items like high fat beef.
Bottom Line
Given the severity of this person, I suggest trying suggestions for 2-3 months and then gets retested. I expect significant changes — but that is likely just a course correction. We need to do more course corrections to get back to a safe harbor.
ALWAYS REVIEW WITH YOUR MEDICAL PROFESSIONAL BEFORE STARTING
%7D%7BN%7D%7D%7D&bg=ffffff&fg=000000&s=0)
Recent Comments