Follow up Microbiome Analysis from a prior post

This is a follow up to my blog post of Dec 30, 2021. Rosacea, Circulation and mild CFS. The person has tried the suggestions, and now we will attempt to see what the consequences are and the next set of suggestions.

Remember, the suggestions are based on mathematical modelling using clinical studies on study populations, so they may work or not work for individuals.

I am a computer scientist and a statistician. I am not licensed to practice medicine, and where I live has strict laws about ‘appearing to practice medicine’. What I can do for readers is to write a public blog (anonymous) from your data and back story as an education post on using the software and the statistics it produces. I cannot consult. The content should be reviewed by a medical professional before implementing.

High Level Measures

Bacteria Selected using Expert Criteria

Use JasonH (15 Criteria)56
Use Medivere (54 Criteria)56
Use Metagenomics (59 Criteria)56
Use Nirvana/CosmosId (36 Criteria)56
 Use XenoGene (22 Criteria)56
 Standard Lab Ranges (+/- 2 Std Dev)86
Box Plot Whisker3027
Kaltoft-Moltrup Normal Ranges7884
Percentile in top or bottom 10 %6399

My impression is that the microbiome has become more diverse, in one sense, unstable. The increase in the number of bacteria types reported (591) moved it just above the typical count for BiomeSight (578). My personal experience is that this is a good sign, the microbiome is changing, I experienced this spike is variety before my microbiome settled down into a new, healthier normal.

Symptoms Change

“This things has improved:

  • Less bloated
  • Seborrhoeic dermatitis is gone 
  • Better stool
  • Better libido

Using the regression for all symptoms we had regressions for, we had 154 improved out of 209 items, or 74% had improvement in the prediction of symptoms.

Overall: Appears to be Improved

This person was a challenge originally because there was no dominate shifts or “smoking guns”. Being at the 99+% for Dr. Jason Hawrelak recommendations and the same items returned from other expert suggestions (many with more criteria) had no significant change. Supporting improvement: Increase in bacteria types closer to typical; significant decrease in number of Unhealthy Bacteria; improved symptoms; and last, prediction of symptoms had a major improvement.

Next Round of Suggestions

After the above sample, he actually started two more items:

  • 10 days with doxycycline
  • started to take rosemary “Feels pretty good taking it. ” 


There are many Ways of Choosing Probiotics, I will look at two below:

KEGG AI Computed Probiotics

The differences actually shocked me, a very very dramatic difference. On the current sample I see what is often on ME/CFS patients list appear at the top: miyarisan (jp) / miyarisan with also L. Plantarum Probiotic Powder. This suggests that he is moving towards/through a more typical ME/CFS microbiome. Given that he has issues but everything appeared normal or good, I take this as a good sign – we are exposing the issues.

Prior SampleCurrent Sample / L. Brevis Probiotic Powder
2 / L. Rhamnosus Probiotic Powder
digestive care
Metabolics / Lactobacillus Rhamnosus Powder
miyarisan (jp) / miyarisan
spain (es) / bivos
spain (es) / kaleidon
spain (es) / ns florabiotic instant
spain (es) / suerobivos
reg’activ / immune & vitality
1.83 / L. Salivarius Probiotic Powder
miyarisan (jp) / miyarisan
enviromedica terraflora sbo probiotic
bio-botanical research / proflora4r restorative probiotic
perfect pass / perfect pass probiotic bacillus spore
730.25 / L. Plantarum Probiotic Powder
goodbelly drink
jarrow formula / ideal bowel support® lp299v®
Metabolics / Lactobacillus Plantarum Powder
Prescript-Assist®/SBO Probiotic
Sun Wave Pharma/Bio Sun Instant

We also have the option of probiotics based on symptoms (adjusted for the microbiome). See Using Samples and Symptoms to Suggest Probiotics post. The data is shown below in decreasing weight order. The nice thing to see is the decrease in the weight of everyone. One totally disappeared (the sole enterococcus faecalis one). It is interesting to note that while above using only the microbiome and resulted in major shifts between samples, when the symptoms are combined the suggestions are very similar and actually reflect improvement of the microbiome.

Prior SampleLatest Sample
Prescript-Assist®/SBO Probiotic (34)
enviromedica terraflora sbo probiotic (4.4)
symbiopharm / symbioflor 1 (3.8)
klaire labs / biospora (3.7)
microbiome labs/ megasporebiotic (3.6)
Prescript-Assist®/SBO Probiotic (22.6)
enviromedica terraflora sbo probiotic (3.5)
perfect pass / perfect pass probiotic bacillus spore (2.9)
klaire labs / biospora (2.8)
microbiome labs/ megasporebiotic (2.7)


I am going to do my current practice of relying on consensus reports because they are now quick to generate. I will be doing a consensus from:

I will be including everything, since the reader is able to persuade his medical professional to prescribe.

Consensus Results

The following are my picks from the options presented. I provided some links to where it helped ME/CFS –i.e. the suggestions are reasonable

Items to reduce or avoid

Remember: These are suggestions, items that improve odds.

Prescription Suggestions

This is done using advance suggestions and flipping the selections:

We do the same 4 methods of picking bacteria above

The top suggestions were:

Secondary positive suggestions are:

Dangers of Filtering

The person tried using the ME/CFS filter and got very different results. This person has mild ME/CFS; the studies on the US National Library of Medicine are for ME/CFS are typically severe and matches a yard of criteria for inclusion in the study. It is often not safe to use there filters when you self-diagnosis or are mild/controlled.

Bacteria Selected using Expert Criteria

MethodFiltering By ME/CFSLatest
Use JasonH (15 Criteria)46
Use Medivere (54 Criteria)46
Use Metagenomics (59 Criteria)46
Use Nirvana/CosmosId (36 Criteria)46
 Use XenoGene (22 Criteria)46
 Standard Lab Ranges (+/- 2 Std Dev)36
Box Plot Whisker527
Kaltoft-Moltrup Normal Ranges1284
Percentile in top or bottom 10 %899
Adding filtering dropped the bacteria consider by a large amount (up to 92%!!).

My usual criteria has been to have at least 1-2 dozen bacteria. With the new consensus report, having a large number of bacteria seems to produce clearer results.

Bottom Line

Suggestions to be discussed with their medical probiotics

In terms of prescription (doing rotation with breaks between):

Supplements to try:

We saw improvements between the sample when this reader implemented some of the suggestions. Remember, the suggestions improves the odds, they do not guarantee nor is there any requirement or protocol to follow.

Follow up Comments from the person

Watched the video you uploaded – it was great to get a video and watch how you did for my test.
Quercetin and resveratrol seems to be something to avoid when I did it your way – so I will cut that out.

High red meat and high beef diet seems to be something to avoid (felt it myself also).Been cutting out all red meat for some days now – feels pretty good.
One thing to avoid is also “vegetable/fruit juice-based diets”. I drink like 1-1,5 liter juice a day. I will try to reduce it but it’s really hard – do not feel good eating to much meat, fat or starch. So when cutting out juice I do not get enough calories. Calories are really important to me – which I also see now on my avoid-list – “low energy diet/ calorie restriction” is on it. 
One other thing that is interesting is that potatoes is on the “Highest Adverse Risk”-list. I eat potatoes every day.

Key Bacteria for Symptoms

I extracted out the items that are recurring as good predictors for various symptoms. These are listed below. This is intended for those interested in research and diving deep. See AI Computed Probiotics from Symptoms for background.

See also Key Enzymes for Symptoms and Key Compounds for Symptoms. What is surprising is that “non-of-the-usual suspects” were included with the exception of spotted fever groups (i.e. Rickettsia, Lyme) and Staphylococcus aureus (see Staphylococcus aureus – the CFS maintainer?)

Thermoanaerobacterales Family III. Incertae Sedisfamily543371
unclassified Clostridialesfamily186813
unclassified Chloroflexinorank167963
Acetonema longumspecies2374
Anaerobranca zavarziniispecies436000
Anaeromyxobacter dehalogenansspecies161493
Bacteroides propionicifaciensspecies392838
Bifidobacterium thermacidophilumspecies246618
Brevibacterium paucivoransspecies170994
Brochothrix thermosphactaspecies2756
Caloramator mitchellensisspecies908809
Candidatus Tammella caduceiaespecies435141
Carboxylicivirga linearisspecies1628157
Casaltella massiliensisspecies938278
Corynebacterium canisspecies679663
Desemzia incertaspecies82801
Desulfarculus baarsiispecies453230
Desulfonispora thiosulfatigenesspecies83661
Desulfovibrio idahonensisspecies575978
Desulfovibrio psychrotoleransspecies415242
Desulfurispirillum alkaliphilumspecies393030
Dolosigranulum pigrumspecies29394
Haliscomenobacter hydrossisspecies2350
Lactobacillus japonicusspecies29399
Marvinbryantia formatexigensspecies168384
Mediterraneibacter faecisspecies592978
Megasphaera hominisspecies159836
Phascolarctobacterium succinatutensspecies626940
Proteiniborus ethanoligenesspecies415015
Pseudoflavonifractor capillosusspecies106588
Pseudoramibacter alactolyticusspecies113287
Pseudoscillatoria coraliispecies693994
Skermanella aerolataspecies393310
Slackia sp. NATTSspecies647703
Spiroplasma ixodetisspecies2141
Staphylococcus aureusspecies1280
Streptomonospora albaspecies183763
Thermoactinomyces vulgarisspecies2026
Thermophagus xiamenensisspecies385682
Thermovenabulum ferriorganovorumspecies159731
Uliginosibacterium gangwonensespecies392736
Veillonella cricetispecies103891
Verrucomicrobium spinosumspecies2736
spotted fever groupspecies group114277

Key Compounds for Symptoms

I extracted out the items that are recurring as predictors for various symptoms. These are listed below. This is intended for those interested in research and diving deep. See AI Computed Probiotics from Symptoms for background

See also: Key Enzymes for Symptoms and Key Compounds for Symptoms

3-Oxo acidC3H3O3R
4-Hydroxybutanoic acidC4H8O3
Amino acidC2H4NO2R
D-Amino acidC2H4NO2R
Geranyl diphosphateC10H20O7P2
N-Acylneuraminate 9-phosphateC10H17NO12PR
Thioredoxin disulfideC10H12N4O4S2R4
Trimethylamine N-oxideC3H9NO

Key Enzymes for Symptoms

I extracted out the items that are recurring as predictors for various symptoms. These are listed below. This is intended for those interested in research and diving deep. See AI Computed Probiotics from Symptoms for background.

See also: Key Bacteria for Symptoms and Key Compounds for Symptoms

EcKeyOtherNameEnzymeName dismutase;superoxidase dismutase;copper-zinc superoxide dismutase;Cu-Zn superoxide dismutase;ferrisuperoxide dismutase;superoxide dismutase I;superoxide dismutase II;SOD;Cu,Zn-SOD;Mn-SOD;Fe-SOD;SODF;SODS;SOD-1;SOD-2;SOD-3;SOD-4;hemocuprein;erythrocuprein;cytocuprein;cuprein;hepatocupreinsuperoxide:superoxide oxidoreductase reductase;CDP-4-keto-6-deoxyglucose reductase;cytidine diphospho-4-keto-6-deoxy-D-glucose reductase;cytidine diphosphate 4-keto-6-deoxy-D-glucose-3-dehydrogenase;CDP-4-keto-deoxy-glucose reductase;CDP-4-keto-6-deoxy-D-glucose-3-dehydrogenase system;NAD(P)H:CDP-4-keto-6-deoxy-D-glucose oxidoreductaseCDP-4-dehydro-3,6-dideoxy-D-glucose:NAD(P)+ 3-oxidoreductase—NADP+ reductase;FPRflavodoxin:NADP+ oxidoreductase dehydrogenase (acetylating);aldehyde dehydrogenase (acylating);ADA;acylating acetaldehyde dehyrogenase;DmpF;BphJacetaldehyde:NAD+ oxidoreductase (CoA-acetylating) dehydrogenase (NAD+);dihydropyrimidine dehydrogenase;dihydrothymine dehydrogenase;pyrimidine reductase;thymine reductase;uracil reductase;dihydrouracil dehydrogenase (NAD+)5,6-dihydropyrimidine:NAD+ oxidoreductase dehydrogenase;AlaDH;L-alanine dehydrogenase;NAD+-linked alanine dehydrogenase;alpha-alanine dehydrogenase;NAD+-dependent alanine dehydrogenase;alanine oxidoreductase;NADH-dependent alanine dehydrogenaseL-alanine:NAD+ oxidoreductase (deaminating) synthase;YkvM;QueF;preQ0 reductase;preQ0 oxidoreductase;7-cyano-7-deazaguanine reductase;queuine synthase (incorrect as queuine is not the product);queuine:NADP+ oxidoreductase (incorrect as queuine is not the product)7-aminomethyl-7-carbaguanine:NADP+ oxidoreductase[formate-C-acetyltransferase]-activating enzyme;PFL activase;PFL-glycine:S-adenosyl-L-methionine H transferase (flavodoxin-oxidizing, S-adenosyl-L-methionine-cleaving);formate acetyltransferase activating enzyme;formate acetyltransferase-glycine dihydroflavodoxin:S-adenosyl-L-methionine oxidoreductase (S-adenosyl-L-methionine cleaving);pyruvate formate-lyase activating enzyme;pyruvate formate-lyase 1 activating enzyme[formate C-acetyltransferase]-glycine dihydroflavodoxin:S-adenosyl-L-methionine oxidoreductase (S-adenosyl-L-methionine cleaving) molybdotransferase;MoeA;Cnx1 (ambiguous)adenylyl-molybdopterin:molybdate molybdate transferase (AMP-forming);glycolaldehydetransferasesedoheptulose-7-phosphate:D-glyceraldehyde-3-phosphate glycolaldehydetransferase;geranyl-diphosphate synthase;prenyltransferase;dimethylallyltransferase;DMAPP:IPP-dimethylallyltransferase;(2E,6E)-farnesyl diphosphate synthetase;diprenyltransferase;geranyl pyrophosphate synthase;geranyl pyrophosphate synthetase;trans-farnesyl pyrophosphate synthetase;dimethylallyl-diphosphate:isopentenyl-diphosphate dimethylallyltranstransferaseprenyl-diphosphate:3-methylbut-3-en-1-yl-diphosphate prenyltranstransferase transaminase;glutamic-oxaloacetic transaminase;glutamic-aspartic transaminase;transaminase A;AAT;AspT;2-oxoglutarate-glutamate aminotransferase;aspartate alpha-ketoglutarate transaminase;aspartate aminotransferase;aspartate-2-oxoglutarate transaminase;aspartic acid aminotransferase;aspartic aminotransferase;aspartyl aminotransferase;AST (ambiguous);glutamate-oxalacetate aminotransferase;glutamate-oxalate transaminase;glutamic-aspartic aminotransferase;glutamic-oxalacetic transaminase;glutamic oxalic transaminase;GOT (enzyme) [ambiguous];L-aspartate transaminase;L-aspartate-alpha-ketoglutarate transaminase;L-aspartate-2-ketoglutarate aminotransferase;L-aspartate-2-oxoglutarate aminotransferase;L-aspartate-2-oxoglutarate-transaminase;L-aspartic aminotransferase;oxaloacetate-aspartate aminotransferase;oxaloacetate transferase;aspartate:2-oxoglutarate aminotransferase;glutamate oxaloacetate transaminaseL-aspartate:2-oxoglutarate aminotransferase sulfurtransferase;thiosulfate cyanide transsulfurase;thiosulfate thiotransferase;rhodanese;rhodanasethiosulfate:cyanide sulfurtransferase selenium transferase;L-selenocysteinyl-tRNASel synthase;L-selenocysteinyl-tRNASec synthase selenocysteine synthase;cysteinyl-tRNASec-selenium transferase;cysteinyl-tRNASec-selenium transferaseselenophosphate:L-seryl-tRNASec selenium transferase sulfohydrolase;sulfoglucosamine sulfamidase;heparin sulfamidase;2-desoxy-D-glucoside-2-sulphamate sulphohydrolase (sulphamate sulphohydrolase)N-sulfo-D-glucosamine sulfohydrolase hydrolase;phosphonatase;2-phosphonoacetylaldehyde phosphonohydrolase2-oxoethylphosphonate phosphonohydrolase;glycogenase;alpha amylase;endoamylase;Taka-amylase A;1,4-alpha-D-glucan glucanohydrolase4-alpha-D-glucan glucanohydrolase;S-adenosylhomocysteine synthase;S-adenosylhomocysteine hydrolase (ambiguous);adenosylhomocysteine hydrolase;S-adenosylhomocysteinase;SAHase;AdoHcyaseS-adenosyl-L-homocysteine hydrolase aminopeptidase;leucine aminopeptidase;leucyl peptidase;peptidase S;cytosol aminopeptidase;cathepsin III;L-leucine aminopeptidase;leucinaminopeptidase;leucinamide aminopeptidase;FTBL proteins;proteinates FTBL;aminopeptidase II;aminopeptidase III;aminopeptidase INULL diphosphatasediphosphate phosphohydrolase 5A hydrolase;CbiGcobalt-precorrin 5A acylhydrolase dehalogenase;2-haloacid dehalogenase[ambiguous];2-haloacid halidohydrolase [ambiguous][ambiguous];2-haloalkanoic acid dehalogenase;2-haloalkanoid acid halidohydrolase;2-halocarboxylic acid dehalogenase II;DL-2-haloacid dehalogenase[ambiguous];L-2-haloacid dehalogenase;L-DEX(S)-2-haloacid halidohydrolase gamma-lyase;homoserine deaminase;homoserine dehydratase;cystine desulfhydrase;cysteine desulfhydrase;gamma-cystathionase;cystathionase;homoserine deaminase-cystathionase;gamma-CTL;cystalysin;cysteine lyase;L-cystathionine cysteine-lyase (deaminating);CGLL-cystathionine cysteine-lyase (deaminating; 2-oxobutanoate-forming) cyclase;adenylylcyclase;adenyl cyclase;3′,5′-cyclic AMP synthetase;ATP diphosphate-lyase (cyclizing)ATP diphosphate-lyase (cyclizing; 3′,5′-cyclic-AMP-forming) 1-methylphosphonate 5-phosphate C-P-lyase;phnJ (gene name)alpha-D-ribose-1-methylphosphonate-5-phosphate C-P-lyase (methane-forming),5-bisthiocarboxylic acid mononucleotide nickel chelatase;LarC;P2TMN nickel chelataseNi(II)-pyridinium-3,5-bisthiocarboxylate mononucleotide nickel-lyase (pyridinium-3,5-bisthiocarboxylate-mononucleotide-forming) racemase;L-alanine racemasealanine racemase dehydratase activating ATPase;archerase;(R)-2-hydroxyacyl-CoA dehydratase activator;(R)-2-hydroxyacyl-CoA dehydratase activase;fldI (gene name);hgdC (gene name);hadI (gene name);lcdC (gene name)reduced flavodoxin:(R)-2-hydroxyacyl-CoA dehydratase electron transferase (ATP-hydrolyzing)—tRNA ligaseL-tyrosine:tRNATyr ligase (AMP-forming)—CoA ligase;acetyl-CoA synthetase;acetyl activating enzyme;acetate thiokinase;acyl-activating enzyme;acetyl coenzyme A synthetase;acetic thiokinase;acetyl CoA ligase;acetyl CoA synthase;acetyl-coenzyme A synthase;short chain fatty acyl-CoA synthetase;short-chain acyl-coenzyme A synthetase;ACSacetate:CoA ligase (AMP-forming)—ammonia ligase;asparagine synthetase;L-asparagine synthetaseL-aspartate:ammonia ligase (AMP-forming) carboxylase;pyruvic carboxylasepyruvate:carbon-dioxide ligase (ADP-forming) ligase (ATP);polydeoxyribonucleotide synthase (ATP);polynucleotide ligase (ambiguous);sealase;DNA repair enzyme (ambiguous);DNA joinase (ambiguous);DNA ligase (ambiguous);deoxyribonucleic ligase (ambiguous);deoxyribonucleate ligase (ambiguous);DNA-joining enzyme (ambiguous);deoxyribonucleic-joining enzyme (ambiguous);deoxyribonucleic acid-joining enzyme (ambiguous);deoxyribonucleic repair enzyme (ambiguous);deoxyribonucleic joinase (ambiguous);deoxyribonucleic acid ligase (ambiguous);deoxyribonucleic acid joinase (ambiguous);deoxyribonucleic acid repair enzyme (ambiguous);poly(deoxyribonucleotide):poly(deoxyribonucleotide) ligase (AMP-forming)poly(deoxyribonucleotide)-3′-hydroxyl:5′-phospho-poly(deoxyribonucleotide) ligase (ATP) chelatase;protoporphyrin IX magnesium-chelatase;protoporphyrin IX Mg-chelatase;magnesium-protoporphyrin IX chelatase;magnesium-protoporphyrin chelatase;magnesium-chelatase;Mg-chelatase;Mg-protoporphyrin IX magnesio-lyaseMg-protoporphyrin IX magnesium-lyase NAD(P)+ transhydrogenase;pntA (gene name);pntB (gene name);NNT (gene name)NADPH:NAD+ oxidoreductase (H+-transporting) reductase (Na+-transporting);Na+-translocating NADH-quinone reductase;Na+-NQRNADH:ubiquinone oxidoreductase (Na+-translocating) phosphate transporter;phosphate ABC transporter;phosphate-transporting ATPase (ambiguous)ATP phosphohydrolase (ABC-type, phosphate-importing) polar-amino-acid transporter;histidine permease;polar-amino-acid-transporting ATPaseATP phosphohydrolase (ABC-type, polar-amino-acid-importing) glycerol 3-phosphate transporter;glycerol-3-phosphate ABC transporter;glycerol-3-phosphate-transporting ATPaseATP phosphohydrolase (ABC-type, sn-glycerol 3-phosphate-importing)

Using Samples and Symptoms to Suggest Probiotics

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).

Where the page is located on the menus

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

Example of suggestions

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:

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:
Where the suggestions without symptoms being taken into account is located here. REMEMBER to raise your display level to see this link
For the same sample as above, we have 2 of the three top suggestions being the same!
  • 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:
lactobacillus salivarius is on an avoid.

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.

Myalgic Encephalomyelitis Microbiome Analysis


I am a computer scientist and a statistician. I am not licensed to practice medicine, and where I live has strict laws about ‘appearing to practice medicine’. What I can do for readers is to write a public blog (anonymous) from your data and back story as an education post on using the software and the statistics it produces. I cannot consult. The content should be reviewed by a medical professional before implementing.

Video walkthru of this blog

Back Story

35yr old, female with ME of around 15 years, Coeliac and Crohn’s diagnosed in 2014 plus simple temporal lobe seizures and endometriosis.

I have tested extremely extensively and most labs are generally normal with the exception of:

  • Prolactin- always slightly above high end of range
  • Lyme ELISA IgM positive but Blot negative (possibly cross-reactive with RF)
  • Rheumatoid factor – one point over upper range
  • LDH – consistently slightly above range
  • Aldolase – tested once, slightly above range
  • ALT – occasionally slightly above range
  • EBV – positive for past infection but never any evidence of reactivation
  • SIBO breath test positive for hydrogen only

I am really struggling with acne… I have been offered Lymecycline for the acne but don’t know if it’s worth the risk. I am currently on a 7-day course of Co-amoxiclav (amoxicillin + clavulanic acid) for an infected cyst (my Biome sample was taken before this).

I was able to control my crohn’s with an elemental diet, followed by strict paleo, then gradually reverting to a more relaxed diet. I did a course of oregano oil which was very harsh on the gut but it got rid of my constant bloating for the first time in my life (this came back and remains since reintroducing carbs). I’m 99.9% sure I’m on the autism spectrum. Interestingly, some of my autistic traits and my ME symptoms abate somewhat for a short period at the beginning of a cold virus (the first couple of days while fever is present). I had a similar temporary reaction to Sulforaphane.

My ME started while I was working full time, following glandular fever in 2006. The onset was characterised by ‘tired but wired’ and rolling PEM, finding it almost impossible to fall asleep until it was nearly time to get up and a complete inability to get into deep sleep. I reduced my working hours gradually but eventually gave up work fully in 2008. Even when ‘healthy’ I never had normal stamina, muscle mass and was really ready for bed by the end of the work day.

Possibly coinciding with my gastrointestinal diagnoses and subsequent avoidance of gluten, my symptoms calmed a bit from the ‘acute’ years, and I began to sleep a little better and get brief symptom-free interludes while at rest. Now, prolonged activity above baseline will lead to a return of the PEM and ‘tired but wired’ symptoms (inability to switch off nonsensical chattering thoughts at night leading to insomnia, feeling hot with chills, feverish, sweats, tossing and turning all night. These symptoms stop on rising but leave me extremely brain fogged from the sleep deprivation, and then repeat the next night. I also experience myalgia, headaches, orthostatic intolerance and strong need to lay horizontal throughout the day. I wake with heavy puffy face and eyelids most days. I go through periods of extreme dry mouth, worse on waking despite always hydrating well.

Even in my symptom-free-at-rest periods, I still struggle to get to sleep some nights, although I keep a strict routine 10pm-10am. I find it very difficult to get into the deeper stages of sleep almost all of the time. Even with all reasonable interventions and ear plugs I am easily startled awake by household sounds and there is usually activity from others from 4-5am onwards.

Microbiome Analysis

I am going to do three levels of Analysis. I will start with generic, then move on to diagnosis using US National Library of Medicine studies and ending with the latest refactor, using bacteria associated with symptoms discovered from uploads to this site.
The purpose of these analysis to get suggestions for the most probable bacteria causing issues.

Overall Health

A domination of unhealthy bacteria
ME/CFS patients tend to have low blood pressure, do ignore hypertension. Acne and Insomnia are reported symptoms — so we have the microbiome matching the conditions. See Q&A for more information.

I will not go step by step (see this post with video for how to do this) but do the following suggestion generation and then look at the consensus

Then I went to Advance Suggestion with Percentile: 15% and the following PubMed conditions (Remember that Display Level must be Intermediate or higher to see this option):

  • Acne – 3 bacteria
  • Crohn’s Disease – 24 bacteria
  • Celiac Disease – 14 bacteria
  • Chronic Fatigue Syndrome – 13 bacteria

Then I went to Symptom Associated to Bacteria (Citizen Science) and found the following applicable items:

For Official Diagnosis: Chronic Fatigue Syndrome (CFS/ME), most of the matched were at the Genus Level

For Official Diagnosis: Chronic Fatigue Syndrome (CFS/ME), most of the matched were at the Genus Level. Most of the matches are usually at the species level.

The result is 40 bacteria added to our hand picked list.

Consensus Highlights

We have a lot of different suggestions over our 10 analysis, often more than 1000!

Suggestions to discuss with your medical professional

On the avoid list:

Also listed was gluten-free diet. I suspect the type of gluten may be significant.

KEGG Probiotics:

The usual one for ME/CFS are at the top (and were also listed in Safer or Safest Takes) :

For supplements:

  • beta-alanine – Percentile: 9.7
  • D-Ribose – Percentile: 4
  • Glycine – Percentile: 13.6
  • L-Cysteine – Percentile: 13.6
  • L-Threonine – Percentile: 6.5
  • Magnesium – Percentile: 0.1

Both D-Ribose and Magnesium are well studied supplements in a ME/CFS context (confirming predictions to clinical practice the items are linked to CFSRemission Blog).

Bottom Line

At this point I will stop for several reasons:

  • Purpose was to show the method
  • Person has brain fog often, so more notes may be counter productive
  • ME/CFS usually have restricted funds, so keeping the number of items low reduces frustrations over not being able to acquire.

A video will be added in the next few days

Questions And Answers

Q: I forgot to mention it, but I do have allergic rhinitis and high cholesterol (as listed under the detected conditions) so this is very accurate.

A: Although I see this often, it keeps surprising me about the accuracy of predictions!

Q: Interesting that it doesn’t pick up on any CFS-related bacteria at a species level, I wonder if this fluctuates and could possibly be because I was relatively low-symptom at the time of the sample, i.e. not in a PEM flare? 

A: You are very likely correct, it does fluctuates. Also, keep in mind that there are many subsets of ME/CFS, so this can be a little hit-and-miss.

Should I take Lymecycline?

A: This is a little complex, we have mixed results.

But when I tried a different antibiotic often prescribed for acne, it is all positive

And my favorite because it reduce inflammation and crosses the blood-brain barrier:

I would suggest making a counter proposal to your medical professional of minocycline instead.

Q: I’d be interested in your opinion on IgY hyperimmune egg powder supplements for targeting gut pathogens

A: There are many retail products like this. What I found are some vet studies[36 listed here]. I found a list of clinical studies here. My general impression is that it is favorable. There is one word of caution, it appears similar to transfer factor — i.e. the IgY may be targeted to specific bacteria/infections only. Hence, my advice would be to buy just one unit of it, if no response, change to a different brand when you finish it. Remember my motto: Rotate, rotate, rotate….

Long term Fibromyalgia with Psoriatic Arthritis


I am a computer scientist and a statistician. I am not licensed to practice medicine, and where I live has strict laws about ‘appearing to practice medicine’. What I can do for readers is to write a public blog (anonymous) from your data and back story as an education post on using the software and the statistics it produces. I cannot consult. The content should be reviewed by a medical professional before implementing.

Back Story

  • Male, 57, very high pain tolerance, little tolerance of irritation (itch, slivers, paper cuts, cold, drafts)
    • Diagnosed with Fibromyalgia at age 35
    • Proposed diagnosis of Psoriatic Arthritis last year
  • Psoriasis (or eczema since 10 yrs old). Usually cold hands and feet.
  • Depression most of my life. 
    • Currently abated by Bupropion, past included: Serzone, Gabapentin,Lyrica
  • Always loose bowel movements, but not diarrhea.  Never constipated.
  • Musculoskeletal pain all my life. Not worse with activity.
  • Restless leg syndrome while sitting,  Had to shift every thirty seconds. 
    • Took pamiprexole for 4 mos.  Stopped PPX, restless leg never returned.
  • Felt sick with Aspartame. (Phenylalanine)
  • About 20 years ago, with the intention of it being an elimination diet, lived on nothing but boiled eggs and (Chinese dish) Beef & Broccoli (with GF soy sauce, sesame oil, garlic, ginger and olive oil). Swallowed Tbsp of turmeric daily.
    • After 7 months, complete remission.
    • Lapsed diet, Remission lasted 4 mos.  Woke up one morning fully sick. 
  • Went on a gluten free, seed free, seed oil free, ketogenic diet about 10 years ago. I suspect the lack of carbs starved a Candida problem.  Much reduced symptoms for about 8 years, downhill slide the last two.
  • Taken arthritis-related herbs the last year.  Researching each, (about a hundred researched, perhaps 15 taken) they all seem to benefit the microbiome.
  • My kidney and liver labs are very good. My heart scan showed zero plaque deposits.
  • Life is pretty miserable.

Note on Fibromyalgia: Bacteria Associated from PubMed, Bacteria Associated from Citizen Science. The only bacteria in common with these two lists is: Hungatella hathewayi. IMHO, Fibromyalgia (as a diagnosis) by itself does not have a strong relationship to the microbiome. There may be strong associations of symptoms to the microbiome.

Microbiome Analysis

I am going to do three levels of Analysis. I will start with generic, then move on to diagnosis using US National Library of Medicine studies and ending with the latest refactor, using bacteria associated with symptoms discovered from uploads to this site.
The purpose of these analysis to get suggestions for the most probable bacteria causing issues

Generic Analysis – Pass #1

This is the suggested path for a first time user to take. We are not targeting for specific issues, rather trying to improve the microbiome towards a typical microbiome. For many people that is sufficient

With the large number of medical issues described above I started by looking at potential medical conditions using National Library of Medicine data. One of the challenges is that the microbiome profiles are often based on naïve patients (i.e. not being treated for anything), he is being treated.

Yellow items are significant risk [i.e. Percentile Greater than (100-prevalence/2) ]

Going further down the page, we see that your microbiome does not have major dysfunction by Dr. Hawrelak criteria (89%ile)

Jumping directly to suggestions

It was interesting that KEGG AI Computed probiotics was very similar to those from a sample that I did a video on [ miyarisan (jp) / miyarisan, enviromedica terraflora sbo probiotic, / L. Plantarum Probiotic Powder, symbiopharm / symbioflor 1]. For supplements, only L-Threonine – was suggested at a Percentile of 6.8%ile.

Expert Criteria with Consensus

I proceeded to the Expert Criteria page and did each option. I then use the consensus report to identify the best candidates.

The first ones (Use…) all just picked 8 bacteria. Standard Lab Ranges picked 12 bacteria, Box Plot: 55, Kaltoft-Moltrup : 73 and Top/Bottom at 10%: 70. As a reminder, the consensus button appears after doing two suggestions sets in 24 hours. An alternative (since the first 5 picked the same bacteria) is to do just one of the “Use” and each of the others, then look at the consensus (left as an exercise to the reader).

Consensus data is kept for 24 hours only, or when a user clear them

The consensus report key suggestions were:

I noted that some of his past diet types are on the Adverse Risk list, i.e. those choices may have contributed to where he is now (not immediate, but keeping to them long term).

National Library of Medicine Conditions – Pass #2

For this pass, we clear our consensus suggestion. We want the suggestions to be specific.

How to clear consensus suggestions so we can build a new set.
We then go to Advance Suggestions.

Since this person has been taking a variety of prescription drugs, I am including that in suggestions. The main reason is to see if there are alternative prescriptions possible that may be more microbiome beneficial. For ME/CFS, often the top items with this choice have been antibiotics that have been used successfully for treating ME/CFS (despite the AI not having that information).

First, we need to determine the conditions that we have data available for and that this reader has. I extracted these as candidates:

I picked the settings below for the first one. I checked ALL types of Bacteria Modifiers.

For each condition, I just change the last select box.

I will skip the avoid list, they should be reviewed by the reader. What is interesting to note was that both antihistamine and antifungal drugs showed up based on the bacteria patterns. This suggests that those two issues may warrant investigation.

P.S. I picked 15% arbitrarily, I like to shoot for an average of 6-10 bacteria per set of suggestions being selected to keep the suggestions focused. You may wish to increase or decrease to tune the number selected.

Using Symptoms – Pass #3

This comes out of this weeks refactor. A video of this feature is below (TO DO).

Again, clear the consensus as we did above. Why, because we will likely be running several list of suggestions.

Where this new feature landed (the name may change a little). Clicking on the link and then entering “Depression” will give you a few choices.
I went with the General Depression — sample size was bigger than the others, and thus the bacteria identified was more.

On the resulting page, you will see checkboxes to pick the bacteria that are likely good candidates to change. I went and checked all of them and then clicked the [Add to Hand Pick Selection] button at the top.

I then checked Fibromyalgia and had no luck (we have a small number of people with this condition, so detection is poor).

I then went on to pain, as shown below, and picked General: Myalgia (pain)

We now have a different list of bacteria

Again, I checked all of the available checkboxes and click to Hand Pick Selection. Then I went to Comorbid: Restless Leg which had only one bacteria with a checkbox.

At this point, I notice that a Hand Picked Bacteria button appears on my samples page

First view what was selected. We have a total of 8 bacteria, sufficient (I hope).

We then pick the suggestions link on the same drop down. Again pick all modifiers.

The list is similar to the early ones. Choline Deficiency means reduce choline intake.
The avoid list — many of the items were seen on the avoid lists above. For a few items we have disagreement, but for most, agreement.

For retail probiotics, we have long list with most having similar benefit

Note why the values are the same, many have nothing in common

Personally, I would likely drop these into rotation (there are no bacteria common to any of them):

  • wakamoto (jp) / wakamoto pharmaceutical intestinal drug
  • / B. Bifidum Probiotic Powder
  •  shin biofermin (jp) /s
  •  optibac / saccharomyces boulardii

Bottom Line

Above we saw three different approaches to obtaining suggestions. There was agreement between each of the approaches for over 70% of the items. My usual suggestion to discuss with your MD before starting:

  • Take 2 of the following probiotics for two weeks and then rotate to a different pair
  • For foods and supplements
    • Niacin keeps appearing as a to take
    • Do NOT take a B-Complex, several of the B Vitamins are counterindicated
    • Barley (suggest for porridge on most days)
    • Inulin
    • For diet type — the reader needs to do a little work to decide what to do. Some of what he has tried are on the avoid list!

“I feel like I am hyper coagulated..”

Someone mentions that to me. My initial response is simple… don’t jump to conclusions without supporting lab results. Without lab results, you may feel the same from any of many condition that produces hypoxia (low oxygen delivery). Hyper-coagulation is one possible cause there are others. This is a quick list of items to get objective measures from..

  • First, get and monitor your saturated oxygen level. A lot of people have the Fingertip Pulse Oximeters. They are cheap and often under $20. With COVID persisting, they should be in every home first aid kit!
    • Both my wife and I wear a watch that records Saturated O2 every 15 minutes and we can view our history on our phone. See this post.
Fingertip Pulse Oximeter, Blood Oxygen Saturation Monitor (SpO2) with Pulse Rate Measurements and Pulse Bar Graph, Portable Digital Reading LED Display, Batteries and Carry Case Included

Bottom line: Get lab results when possible and do not speculate. I recall that some supplements will improve certain types of hypercoagulation and make other types worse. Act from objective knowledge and not speculation.

No description available.
Example of our family testing. This was done by Hemex, a specialized lab in coagulation. Any one of these can indicate hypercoagulation. Many physician will order only one or two items and them proclaim there is no evidence…

A reader experience with CFS and Autism

Sue, a reader in Australia, shared her experience below on the challenge of taking supplements. Other people may have the same challenge. In some cases, retail products additives can be the source of problems. In this house, we tend to make our own or use additive free.

The problem began when Ken Lassesen’s brilliant AI came up with oils my daughter must take by swallowing them, with high probability of success. Coriander, thyme, lavender, perilla oils. She’s been sick to the point of house-bound for years, and we’ve got to get her into life. She has autism and chronic fatigue, and since 2018  Ken Lassesen’s AI  has at least got her out of bed (unexpectedly, skads of licorice and thyme leaves are  key players here) when no kindly, well-informed, hard-working doctor could achieve that.

So- where to get the oils? First call was to my excellent local pharmacy in Sydney – Newton’s – and they told me that their oils shouldn’t be ingested, but they’d heard of a place “somewhere in NSW” called “something like” TERRA,  and they were “a bit expensive”. I found and rang TERRA, and they sent me the oils, one expensive from a company that’s licensed to provide ingestible oils, the other from a second company that I was assured was as good but hasn’t gone through the rigours of getting a license.

Now, how to get them into her. We tried dripping it into Bonvit gel caps  from our local chemist with some “blotting paper” in the bottom of thiamine which she has to take anyway, but they disintegrated almost as we looked at them. Bonvit are  fine with powders, but oils aren’t powders! Then we tried Surgipack’s capsules which were sturdier so they  lasted until she put them in her mouth. A few seconds later, howls of pain.  

Next morning we thought we had the solution and tracked down two sizes of Surgipack, so we put the oil and thiamine mixture into the smaller one and put that inside the bigger one, carefully wiping down the outside surface. We were very pleased with ourselves until a phone call with howls of pain with a burning oesophagus and stomach. I talked her into trying to take them right in the middle of breakfast, and she agreed to give them another go. It was still painful afterwards but she bravely soldiered on, saying that “it might be working” !  But after five days, the pain got too much to bear and she said she’d  “never have that stuff again. However good it’s supposed to be”.

But there must be nasty meds taken all the time, so how do the commercial companies get them into people without the whole nation howling in pain, with mass revolts? They must know something we don’t. 

We have a lot of private compounding chemists here so I rang around and asked them, one after another, if they used special capsules and could I buy some. But they make their money out of compounding, not selling their ingredients, so no. At last I chanced on a chatty girl who said what I needed was enteric-coated capsules”, but that her company couldn’t supply them. She vaguely mentioned legal reasons.

So then, the internet. We immediately found a supplier of enteric capsules in Australia, The Capsule Guy, costing $17 for 250 capsules. They come in sizes and we chose smallish ones, so we could put them inside a larger Surgipack capsule in case of dribbles.

We began 4 days ago. No howls. Then joy.  Yesterday afternoon, her birthday, she went out to the party of a childhood friend who has the same birthday. She only stayed 3 hours before she wilted, but she went out and we are over the moon. Thank you, Ken. You are bringing our daughter into life.


Reading an account like this makes all of the hours that I spent on the blog and web site worthwhile. Thank you Sue for sharing! P.S. Sue started in 2018, it is not an overnight turnaround, it is a slow long march. Each person is unique, as is their microbiome. Microbiome Prescription is specific to an individual’s microbiome and not their diagnosis.

Distribution of Jason Hawrelak Criteria

I offer Jason Hawrelak Criteria on my health page on Microbiome Prescription as shown below.

There are 15 Criteria

Doing the refactor, audit and re-validation of the site, a reader asked about their percentile ranking against the above. Good question and relatively easy to answer.

Percent of Healthy MatchesCountSamplesCommutive Percentile

Bottom line is simple, if you have 40% (6 items) healthy – you are better than typical person. I would not be concerned about trying to raise it higher…. Not a single person has made it over 80% of the criteria….

Plotting the data revealed a logistic curve, which seems to constantly occur with microbiome data.