End Products and Autism, etc

Many months ago, I extracted information from various sources on what end products/metabolites are produced by various bacteria. My reasoning was simple: the symptom may be due to too much OR too little of some metabolite/chemical, like lactic acid. Lactic acid is produced by many different bacteria. If we try looking for patterns with the end products then the specific bacteria become less important, and the combined results of many taxa become apparent.

Over the last weekend, I attend a conference on Autism and the speakers mentioned finding some metabolites too high or too low. I decided to revisit the end product pages and added some new pages. My focus is autism but the approach in this post can be applied on other matters.

The quick summary page

This page just lists the symptoms (like autism, POTS, FM, ME/CFS) and how many metabolites appear to be statistically significant. I should mention that often the pattern is not autism has the lowest values or highest values, but often land in the high medium range — missing high values and low values.


Looking at AGE 60-70, you can see a lot of age related changes! Lower vitamin B12 and butyrate, higher hydrogen sulfide

Age: 60-70, http://microbiomeprescription.azurewebsites.net/Data/EndProductExplorer2?includes=192

My original thinking is that by taking select supplements you can bring up (or take down) the levels in the gut. My rationale is simple, end products are used by other bacteria. If you are short on one, it will alter the microbiome balance across many taxa, a cascading effect.

Autism: http://microbiomeprescription.azurewebsites.net/Data/EndProductExplorer2?includes=262

With autism, looking at the numbers above there is a pattern of “not in the middle”. The amounts being produced are too high OR too low often. Some items are in one direction:

  • Formic Acid: Too high

We also have finer resolution available which (a surprise to me) often find more items, as shown below.

When we go to finer levels, we actually get more stronger relationships

For Lactic Acid, there are a number that are sky-high. A similar pattern is also seen with a subset of ME/CFS. Trytophan has the values at the highest and lowest 6%. Etc

Bottom Line

Unlike taxa where diet and supplements modifications are obvious steps, a lot of eurekas here may need physicians to know how to proceed. A few items like low production of Vitamin B12 and some amino acids have over the counter supplements available.

Child Autism microbiome over time – Part 2

In Child Autism microbiome over time – Part 1 , I reviewed the predicted symptoms for each samples and saw a regular pattern. As a result, I wrote some code to cluster the taxa from most common predicted symptoms… in english, identified the bacteria that appears to be causing most of the regular predictions.

To get to this new report, on the samples page, click [Compare Samples to Each Other]

On the next page, select the samples to be included and click [Probable Symptoms Cluster]

The next page does the analysis automatically, showing the common symptoms and then ranking the bacteria matching these symptoms

Just check the ones that you wish to include. My rule of thumb is to go down to 1/2 of the highest value (i.e. 16.5 or higher, since the high is 33). Then create the custom profile. At this point it will evaluate all of the samples and add this hand picked to the best sample. (You may wish to delete other hand picked taxas before doing this — so you can find it!)

The handpicked list is in gold.

You can view the choices if you wish

At this point, we are ready to try suggestions

What happens here is not unexpected. Because we are dealing with a different selection of bacteria, suggestions may contradict, others in agreement with Child Autism microbiome over time – Part 1

There is no easy solution.

  • Items recommended by both or avoid by both are the easy part.
  • When there is disagreement, I usually opt to omit them unless there is a specific reason to include.
  • If it is one one but not the other, I am inclined to go with the suggestion.


This is artificial intelligence and that usually mean that it can be tuned or adjusted. Whether it makes the results better or worst may be subjective.

Raising the value reduces the symptoms, reduce the bacteria identified
Lowering, increases the symptoms and the number of taxa

A Second Child

There is another mother with a boy with autism who has also been doing regular samples. To my surprise (and likely my AI’s delight!!!) we came up with similar results for symptoms

Bottom Line

This will likely work with many different conditions. I tried it for myself across all of my samples (in remission of ME/CFS and in relapse) as shown below… Needless to say, the AI did an awesome job!

We do not want to work from a single sample. We want to take regular samples to quiet down random noise in the microbiome and identify the masterminds behind any issues.

Getting Started

If you are new, use Thryve Inside (there is a Gut Club Discount Code ) because it is the most used and thus we have more data on it.

Child Autism microbiome over time – Part 1

A reader granted permission to review their child with autism microbiome over time, especially in light of my recent post Technical Study on Autism Microbiome. It presents an opportunity to better understand the dynamics of the microbiome with autism. I say dynamic, because the microbiome keeps changing — sometimes in minor ways, other times major. It is a moving target. This is the first part of several posts. There is a lot of data and some of it caused me to write new tools to answer some questions that arose. I believe that microbiome has significant impact on the severity of symptoms in autism.

One time microbiome testing is foolishness… you need regular ongoing testing to discover what hangs around and what is a visitor passing thru… What hangs around is what is important… Visitors can often be ignored. The microbiome is very dynamic.

You don’t make a call on who going to win the next election by asking the first person you meet on the bus

Ken Lassesen

Timelines of Key Bacteria Taxa

Presentation of the material


I find that Species reported often and be informative. Conventional wisdom is that more is better… There was a dramatic change in early 2019.

“Sample after probiotic L.reuterei and camel milk. Consistency soft solid
Diet: consisted of vegetable soups and also had introduced celery juice in the mornings”

Consensus Taxa

Prior post, Technical Study on Autism Microbiome cites:

The values are opposite – not high but very low.

Blautia being Low

Low values are the norm, often at the very bottom

Bifidobacterium being High (in some Published studies)

Usually low, but with two sudden spikes to high. This is reported low in some published studies and high in other studies. It may a volatile taxa with autism — I have seen major swings in another child with autism of Bifidobacterium.

Lactobacillus being High (in some Published studies)

Almost follows Bifidobacterium in flipping between low and high

Citizen Science Taxa

From most significant downwards..

Erysipelatoclostridium genus / Erysipelotrichia class / Erysipelotrichales order – Low

Clear match of pattern

Veillonella – High

Again agreement, with an increasing over time pattern

Senegalimassilia – High

With recent swings from none to high

Marvinbryantia – Low

Desulfovibrionales (Order) Low. This is reported in 92% of this lab’s sample. 6/11 having none is a low probability event.

Eggerthellales (Order) Low

Intestinimonas – Low

Was low and this year jumped

Anaerotruncus Low

Some random jumps but usually low

Pseudobutyrivibrio Low. This one is only seen 57% of the time in Thryve (which is what was used here) and 98% in uBiome. This distribution is not a very rare/unusual one like some following.

Burkholderiales Unusual!! Low and High (few middle ranges). Clicking the link I see 94% have measurable quantities – for only 1 out of 11 samples to have it is a 1 in 174,000,000,000 chance….

Deltaproteobacteria Low

Very high for a while and then collapse to nothing

Alistipes – Low. Clicking thru I see 85% have measurable quantities. For 7/11 samples having none (and when it does, very low)… suggests it’s absence may be significant.


Borderline for Significance

Terrisporobacter Medium Low. Clicking this link I see 75% of samples have measurable amounts. For 10/11 samples to have none suggests some significance to its absence.

Breaks from the pattern

Summary Line #1

The following bacteria are rarely seen at all with this child but is very common in other samples. These are also reported as low across the Autism spectrum from Citizen Science:

The premise that we are working off it that the unusual is contributing to autism.

A good question to ask – Is this familial?

The mother has also done a Thryve sample so comparison was easy

Suggesting that it is unlikely due to DNA. It does suggests that giving the daughter a lot of kisses (especially on hands before meals) may have some benefits.. 😉

Suggestions #1: Dive down on this oddity

I went to Bacteria Symptom Explorer Plus (Autism via Citizen Science is there). And custom picked these 4 bacteria taxa that we want to increase. I then create a [Hand Picked Taxa Suggestion].

Our four taxa – all of these are low

We asked for 30 items, and got less – running with direct citations. It was interesting to see Triphala on the list because it reduces many bacteria and may as a consequence increase these!

Note that certain species of Bacillus and Bifidobacterium are good and others are bad!

I tried various ways of expanding suggestions and found only parents increased the list slightly. ß-glucan + linseed(flaxseed) + high fruit intake may translate to Iron-fortified Oat or Barley porridge with flaxseed and fruit for breakfast. Supper with Oregano, Turmeric with lots of cruciferous vegetables (broccoli cabbage)

Looking back at gut based on Suggestions

We look at the probiotics suggested and levels that this person have of then.

Thrive does not report on this one
Nor on this one
Nor this one
Eureka — we have one reporting
It comes and goes — mostly zero. Taking it makes sense
At the unspecified genus level this is a do not take a generic mixture of bifidobacterium

Predicted Symptoms

Going thru the samples, using 0.6 as the cut off point for predictions. Fatigue may manifest itself as irritability. Every single microbiome sample had autism as #1 predicted symptom.

The ouch!

The mother’s microbiome had a surprise… a weak autism-like profile. Only 12 matches (the daughter ranged from 14-21). The mother when she first emailed gave a “no health issue” description.

” Btw, I looked at the probable symptoms for my own sample and .. [many] were right on specially the neurocognitive ones… but a shocker at the blood type.. is 100%   Although none of the symptoms listed have ever been a concern, so sort of lived with it.. but now I’m so curious to exploring the data more.Thank you again for building this awesome tool!”

– Mother

This weekend I attended a conference with two presentations by Jason Hawrelak on Autism and the Microbiome. He presented his hypothesis that with modern western life, each generation’s microbiome becomes a subset of their parent’s microbiome. As a general concept, I agree if there are no radical changes of lifestyle (inconvenient changes usually). It’s a rational explanation for the increase of autism and other microbiome associated conditions. With this model, the mother was likely on the path towards autism(which was likely delivered to her by her mother) and with an additional iteration subsetting her microbiome… her child was dropped into it.

I believe that it is possible to recover significant amount of the lost microbiome. A simple first step is to spend weekends working on an organic farm as volunteer labor. If the kid eats dirt, or sucks on grass or wheat on this farm… he may be potentially repopulating some of the microbiome. I recall walking with my father (a farmer) and his picking straw and grass for me to suck on (unwashed) — I was getting hay bacillus or grass bacillus, a.k.a. Bacillus subtilis. “Farmer common sense medicine”

Next Installment

Coming next is Suggestions #2, looking at the bacteria that dominant the prediction of symptoms across the many samples, namely

  • Official Diagnosis: Autism
  • Comorbid: Constipation and Explosions (not diarrohea)
  • Official Diagnosis: Mast Cell Dysfunction

With these bacteria

NameRankTimes Cited
Dorea formicigeneransspecies33
Dorea longicatenaspecies26

Please remember, I am not a medical professional. I am a professional statistician, artificial intelligence engineer and software developer (Microsoft, Amazon, Starbucks etc). I extracted “facts” from medical literature and use these facts to drive a fuzzy logic inference engine (commonly known as Artificial Intelligence).

The intent is to explore logical possibilities that may warrant future studies by medical professionals using statistics.

Human Coronavirus – the Facts 😷

I happen to live in the next county/parish over from where there is 2019-nCoV patient confirmed by CDC. This describes symptoms for under 60’s

My site with forecasts. The forecasts have been very close to actuals (often too low).
US was at 6,500 later that day
New Chart Type added. Many countries are < 1%, i.e. little control

These are some basic notes…..

Denmark total shutdown, closed borders appear to be working

I have stopped updating these daily, go to http://microbiomeprescription.azurewebsites.net/covid/chart for the latest with forecasts.

The Epidemiological Characteristics of an Outbreak of 2019 Novel Coronavirus Diseases (COVID-19) –

CNN Daily Press Summaries

While there are a lot of stories, often > 100 per day. A read thru is very informative

Some probably correct information

Mutations are seen

Corona virus are well known to mutate easily, see this article on Scientific American.

The Mug Shots

Coronaviruses. From CDC/ Dr. Fred Murphy

First 4 virus have a 3% Mortality, except for Cancer patients where it is 24% [2019]

  1. HCoV 229E (alpha coronavirus) –  common cold
  2. HCoV NL63 (alpha coronavirus)
  3. HCoV OC43 (beta coronavirus) –  common cold
  4. HCoV HKU1 (beta coronavirus)
  5. MERS-CoV (the beta coronavirus that causes Middle East Respiratory Syndrome, or MERS)
    1. 851 dead in 2468 laboratory-confirmed cases[2019]: 34.5% Mortality
  6. SARS-CoV (the beta coronavirus that causes severe acute respiratory syndrome, or SARS)
    1. 774 dead in 8099: 9.6% Mortality
  7. 2019 Novel Coronavirus (2019-nCoV)
  8. “China’s mystery new coronavirus has so far infected at least 555 people and killed 17, ” Jan 22 Mortality rate is based on those that recover. This means that Mortality is at least 3% (assuming no one else dies). Actual mortality may be easily 20+% or more.


Several high ranking officials in the US and China cite that it will “die down”. Since we are dealing with a cold virus, we have past studies available. Yes – the infection rate will go down, but instead of 100 dead per day, we may be talking just 40 dead per day during the summer.

The effect of hydrotherapy on the incidence of common cold episodes in children: A randomised clinical trial (2003)

Deaths etc

See World Map of Cases maintained by John Hopkins University


The falling mortality may be due to the increasing use of various anti-viral drugs: ” A 35-year-old Seattle-area man identified as the nation’s first person with coronavirus developed pneumonia while in the hospital but has improved after getting antiviral medication, according to a new study. ” [Src] Thailand: anti-HIV and anti-influenza drugs [src]

And the Wuhan Municipal Health Commission also said on Monday that at least 15 health workers in Wuhan have been infected with the virus, with one of those workers put in a critical condition. Jan 20

No photo description available.

Scientists have revealed each infected person is passing the virus onto between 1.4 and 2.5 people. It is known as the virus’s basic reproduction number – anything higher than 1.0 means it’s self-sustaining. .. The fact only 25% of reported cases are severe is a mixed blessing. Yes, that is less dangerous than Sars, but if those hard-to-detect mild or maybe symptomless cases are contagious too, then it is much harder to contain.


Incubation Period

This is the time from meeting someone with the virus and getting it until you show symptoms.

  • “the incubation period is approximately 2–14 days; median incubation period is slightly more than 5 days.  ” CDC on MERS
  • “The incubation period for SARS is typically 2 to 7 days, although in some cases it may be as long as 10 days. In a very small proportion of cases, incubation periods of up to 14 days have been reported. ” CDC on SARS

Example: Patient X arrives by airline on January 1st, infecting 10 fellow passengers. January 6th, five are showing symptoms. January 15th, we have 11 patients with symptoms. Assuming each infects 5 more, we have 65 patients by the end of the month. 330 by Valentine’s day, 1600 by the end of February.

“[they will not seek treatment] if they believe they merely have a common cold. In fact, previous iterations of the coronavirus are very similar to a common cold. ” New Scientist

Shedding Duration

This is the time between symptoms showing until the person is no longer likely to infect others.

  • Positive rates peaked at 6-11 days after onset of illness for nasopharyngeal aspirates … and 9-14 days for faeces  [2004]
  • The median duration of shedding was 14 days (4–60 days), and 17 of 42 patients had prolonged shedding (≥21 days). Among 31 available nasal samples, 35% were OC43, 32% were NL63, 19% were HKU1, and 13% were 229E. The median shedding duration of HCoV in nasal samples did not differ between strains. [2017] This is for the common cold varieties.

“The common cold is infectious from a few days before your symptoms appear until all of the symptoms are gone. Most people will be infectious for around 2 weeks. Symptoms are usually worse during the first 2 to 3 days, and this is when you’re most likely to spread the virus.”

UK’s NHS (equivalent to US CDC)

“Third US coronavirus case confirmed in California; infection can spread before symptoms showUSA Today Jan 26 Excuse me, the moment it was identified as a coronavirus that should have been assumed!

Mutates Easily

” Bovine coronavirus and canine respiratory coronavirus diverged from a common ancestor in 1951.[34] Bovine coronavirus and human coronavirus OC43 diverged in 1899. Bovine coronavirus diverged from the equine coronavirus species at the end of the 18th century. ” [wikipedia]

China is warning that the mysterious Wuhan coronavirus is mutating, meaning it could spread further and become harder to control. ” [Jan 22]


Dr. Mark Loeb, an infectious disease specialist at McMaster University in Hamilton, Ontario, said a study during an outbreak of the SARS coronavirus found that any type of protection — whether a mask or a respirator — reduced the risk of infections in health care workers by about 85 percent. [src]

  • We estimate that if the entire cohort had used masks consistently, SARS risk would have been reduced from 6% to 1.4% per shift. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3322898/
  • ” At Hospital A, the risk for developing SARS was 12.6 times higher in individuals not using a mask than in those using a mask.  ” https://www.ncbi.nlm.nih.gov/pubmed/18806349
  • A case-control study with 36 cases and 50 controls was conducted of factors associated with the transmission of SARS within the hospital. In univariate analysis, contact with respiratory secretions elevated the odds ratio to 6.9 (95 % CI 1.4-34.6, P= 0.02). Protection was conferred by hand washing (OR 0.06, 95% CI 0.007-0.5, P=0.03) and wearing of N95 masks (OR 0.1, 95% CI 0.03-0.4, P=0.001).  https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2870165/

From Do N95 respirators provide 95% protection level against airborne viruses, and how adequate are surgical masks? 2006 😷

  • “2 surgical masks, which originated from the same manufacturer, showed tremendously different penetration levels of the MS2 virions: 20.5% and 84.5%” i.e. up to 85% of virus particles will make it thru!
  • ” The N95 filtering face piece respirators may not provide the expected protection level against small virions. ( the particle size range of 10 to 80 nm)”
    • Note: “The virion shape is spherical, with an average size of 125 nm [src]. ” so N95 should provide advertised protection.

Approach #1 wear masks always outside of home, #2 home prepared food (wrapped) or prepackaged food heated wash hands before meals. Buy packaged food or food that can be heavily washed. No restaurant or take-away food. When you get home, take clothes off outside (so underwear and mask), then straight to shower (especially work on hair) and new clothes is route #1… external clothes are handled with gloves and go directly into wash each time.

Misc Items

The following may or may not be relevant

At what point do you lock down a City

” Huanggang on Thursday became the second Chinese city to go into lockdown over the coronavirus outbreak that has killed 17 people and infected more than 630. Huanggang had reported 12 cases of the coronavirus as of the end of Monday. ” [Jan 23] The Chinese is applying lessons learnt from the SARS outbreak.

Rational way to Constain it

It is simple, and already happening in China. Make it a legal requirement to wear a protective mask in public.

Consider this scenario: a 22 y.o. Barista at a Starbucks has it and symptoms do not show up for 7 days. During those 7 days, she is shedding virus on to customer’s coffee. She develop very mild symptoms (typical for that age) and continue to work. She may continue to shed virus onto customers for 2 more weeks. If this is a busy Starbucks, we are talking thousands of potential infections.

Bottom Time Lines

  • From contact to Infection
    • 2-14 days, average 5 days
  • From contact until no longer infectious
    • 6 – 74 days, average 19 days
  • Time from last reported case recovered until “all clear”
    • At least 3 weeks, up to 10 weeks.

If the 35% increase of cases per day keeps up (i.e. governments not taking radical action SOON)… on March 17th, 2020. Everyone in the world will have gotten the infection – just simple mathematics.. 2000 * (1.35 ^50) = 8 billion people.

The latest case to be confirmed is in the German state of Bavaria, only the fourth so far in Europe.

It involves a man who contracted the virus at a training event from a visiting Chinese colleague who reportedly did not show symptoms until she returned to China days later.”


THIS is statistically why I advocate that a rational approach is to shut down travel (planes especially, trains, etc) world wide. If regions are free for 3 weeks. Travel between those regions can resume. If you have just 1 case… you are shut down until that case resolves (dead or recovered) plus three weeks…

Supply List

Someone asked about a supply list. My suggestions would be:

  • Prophylactic
  • Masks
    • N95 – sufficient for 3 months, or
    • Powered Air-Purifying Respirator (PAPR) with spare filters
  • Gloves
    • 2-4 boxes
  • Emergency food supply (long life)
    • In case of distribution disruption

The ALA Herb: Shiso

A reader continues going thru traditional folk medicine looking for items off the beaten path that may be of interest and sending candidates to me.

The latest request for me to research is a gem. Mast cell inhibitor, anti-allergy, reduces gut inflammation, likely help with SIBO, etc

EidolonGreen [China Medicinal Herb] Perilla Frutescens (Perillae Folium/Zisuye/紫苏叶/자소엽) Dried Loose Leaves 3 Oz (88 g)

Perilla frutescens var. crispa, or shiso containing 58-65% alpha-linolenic acid [2006] [2008], It is sometimes referred to as purple mint

Microbiome Impact

Bottom Line

For additional information, see https://content.selfdecode.com/perilla-oil/

Needless to say, it will be added to the database at http://microbiomeprescription.azurewebsites.net/ and increased the number of direct citations.