UA-45667900-1

Friday, 30 May 2025

Presume competence? Presume incompetence? Or just stretch boundaries?



I had a strange experience recently: for the very first time, a stranger asked me if my 21-year-old son has autism.

We were on holiday in Northern Spain, staying at a tiny hotel with just 6 rooms. The hotel was in a tiny village near San Sebastian that happened to be on the Camino del Norte, or Northern Way, which is a well known hiking route. The route was originally a pilgrimage to the city of Santiago de Compostella. The route seems to be popular with older Americans and British.

At breakfast there was one long table and so strangers were almost inevitably going to talk to each other. For Americans this is normal behaviour, but much less so for Europeans.

Monty was wearing a new replacement set of wireless headphones, which we had not quite figured out how to reliably connect to all his devices. As a result, he did look like the classic person with autism  wearing ear defenders and looking anxious.

“Excuse me, is he autistic?”, queried a guest who was walking the Camino del Norte with a friend.

It turned out that this fellow guest at the hotel had an adult son with autism, now in his 30s. His son is still doing ABA and does not get taken on holidays. “I'm impressed you take him with you,” he commented.

We felt it necessary to explain that Monty can do a lot: he completed mainstream school, passed his exams like the typical students, and now travels alone by public transport to “work” twice a week. He has been to China, Japan, South America and most of Europe. He can ski down black slopes, play the piano …

Our fellow guest told us how once his son had eloped and a police helicopter had been needed to find him, not surprisingly near an expanse of water. He did not attend school, due to his sensory issues.

This got me thinking about how we presume competence, or indeed incompetence.

These concepts have become quite a topic in the field of inclusive education. They have been rather stretched by the DEI people, but they are worth evaluating.

Most people assume that a person who behaves typically and is fully verbal must have full mental competence. We are surprised when that assumption proves false. For instance, Harvard University has introduced remedial math classes for some students; you wonder how that is possible. Similarly, some high school students in the U.S. cannot read analog clocks. Social media is awash with videos of young adults asking high school aged kids basic questions like "what is 33 divided by 3?" and having them unable to even make a reasonable guess. There is even a meme of teenage American girls being asked "in what country is Alaska?" and one answers Mississippi.

When we see a person who is not fully fluent verbally, most people tend to presume incompetence.

Last night, as Monty and I were completing our evening uphill “rucking” (fast walking with weights), I decided to check something. An out of breath Peter said “Mont, what is 33 divided by 3?” Without hesitation, he replied “eleven.” OK, I can use that example.   

   

Stretching boundaries

I did explain in Spain how we got to the point of travel independence. It was a step-by-step process and did not happen overnight, or by itself. I was asked how far away I was during this process. I did explain that with modern GPS tracking available on phones and air tags, it is now very much safer and easier. But things can and will go wrong – that is life. People learn by making mistakes – best make small ones, whilst you are still young!

I am a proponent of constantly stretching boundaries on the basis that taking many small steps forward can take you a long way. Just as it does for those older folk walking along the Camino del Norte.

Constantly stretching boundaries and gradually extending your comfort zone seems a good approach to autism. 

 

Mission Impossible in 4D

Monty’s big brother took me and Monty to see the new Mission Impossible film last week.  As we were about to buy the tickets, big brother said “Oh no, it's in 4D”. Watching movies in 4D is like being on a plane in severe turbulence. “No problem, he will enjoy it” was my response.

This was an example of presuming competence.

It was a great film to see in 4D, it really is a compelling experience. 100 times better than films in 3D.

Monty loved it.

 

What about those who are never competent?

For the DEI (i.e. not realistic) version of competence, here is a link to the TACA site.

 

Presuming Competence in Autism

Presuming competence means valuing all people, including those with autism, as whole individuals with the right to express their thoughts, feelings, and opinions. For individuals with autism, this includes the right to communicate, the right to be treated their age, to have their views and feelings respected, and to be involved in decisions about their lives, large or small. This article covers ways and things to consider when presuming competence in your loved one with autism. 

Speak Directly to the Person and in an Age-Appropriate Manner

Presume that everyone can understand what is being said.

Do not talk down to people with autism.

Do not use baby talk or a baby voice. 

Etc … 


The problem is that some people have impaired cognition, not just impaired verbal communications skills. They may never be able to safely cross a road independently, and some will grow up to be like a toddler in an adult’s body.

However, some young children diagnosed with level 3 autism have made such great strides in the early years that they have left their greatest challenges behind them. They should no longer be considered at level 3.

The Lancet Commission has wisely stated that you need to wait until the age of 8 before you can reliably diagnose profound autism. For these children, stretching boundaries seems a better and safer approach than presuming competence.

  




31 comments:

  1. This is truly inspiring, Peter. You're an incredible human being and a devoted father. The love and dedication you’ve shown brought a tear to my eye. Thank you for sharing this journey with all of us !

    ReplyDelete
  2. You have a big heart to share your personal experiences with all of us. Thank you. Your blog has helped me last 3 years.

    ReplyDelete
  3. We often get strange questions. I navigate such situations with the following thought: there are two parallel realities that collapsed at that time, mine and an opponent with a question or reaction. I could get him to my and my daughter's storybook starting to process and change something in his mind, or I just flow to my reality and will meet next people who I accept to my journey and memories. This is my choice of what type reality I would like to create and process around myself. Normally I just ignore those noise.
    We try to give our daughter the same types of everyday activities that are appropriate for her motor skills, just like any neurotypical child. Traveling abroad, cinema, entertainment etc.
    Personally, I find it weird to send an autistic child to a facility for the weekend, supposedly just to spend more time with another sibling. Or drop them off during the holidays. Yet, this seems to be typical practice across Europe.
    We visited a developmental pediatrician, and at the end of the appointment, her conclusion was to give us the address of a special needs institution where I could "drop off" my child for the weekend. Welcome to Switzerland. Traveling abroad seems easier than navigating local systems. Presume competence is what we , parents, must advocate in our everyday life.

    ReplyDelete
  4. Hello Peter,

    What do you think of rapamycin (sirolimus) for ASD, have you ever come across anyone use this successfully?

    This could work really well. The drawback is that it would potentially require daily dosing compared to the once weekly dosing that is mainly being used by the longevity or ME/CFS community that are currently running trials for it.

    In post mortem exams of brains of autistic individuals it seems almost all have some sort of lack of synaptic pruning, big question is does mTOR always play a part of this. Lowering mTOR may reverse microglia activation (neuroinflammation) and also clear 'debris' and such like that one researcher mentioned over 10 years ago here: https://www.cuimc.columbia.edu/news/children-autism-have-extra-synapses-brain

    ReplyDelete
    Replies
    1. Yes, some people have trialed rapamycin for autism. They key issue seems to be how well tolerated the drug is - some people do get side effects, meaning they cannot use it long term.

      Delete
    2. There is also everolimus which has a shorter half life and is thoight to pass the blood brain barrier better than sirolimus. Still comes with the same immunosuppressive effects but for those that it works for ot could be the closest thing to a 'cure'/treatment.

      Delete
    3. I've done a trial in the past. It does work but it comes with significant side effects. The immunosuppression is no joke. I've been doing more target therapies now. Right now fucoidan is in my top 3. Stops the downstream effects of p-cresol and increase s tight junctions.

      Fucoidan Rescues p-Cresol-Induced Cellular Senescence in Mesenchymal Stem Cells via FAK-Akt-TWIST Axis

      https://pubmed.ncbi.nlm.nih.gov/29642406/

      Delete
    4. Stephen , what dose of rapamycin did you use and how often? And when you say it worked do you mean that if you were able to continue it you think it might have given even more benefits?

      It seems it needs to be used daily for the correct effects but perhaps a high dose once weekly like how the longevity community uses it might still work without the immunosuppressive side effects

      Delete
    5. Just looked at my notes, 1mg daily for two weeks straight. There was an increase in words and he started to independently use the potty. 40 independent mands at the ABA center. Then he got really sick so I stopped. I then tried 0.5mg for a week similar response. Occasionally I do the longevity hack way but its not as good. Rapamycin does work however I like omalizumab a lot more.

      Delete
    6. Hello Stephen,

      If your son got sick at 0.5mg a day it might be viral reactivation. Rapamycin helps decrease viral load when taken weekly long term but when taken daily there are reports for people that it reactivated their latent EBV or CMV. Taking it with Valtrex or even natural antivirals (monolaurin, lysine high dose etc.) may help but I have yet to try it but am going to soon.

      Delete
    7. Good to know. I've also been using Fucoidan and have been very happy with that too. He has an ENT appointment for his tonsils because he snores and I believe they are enlarged because of his poor immune system. I wonder if valtrex would help decrease the size of them.

      Delete
    8. Chronically enlarged tonsils and adenoids past a certain age could actually be a sign of overactive immune system as well. So if the person gets the cold/sick often compared to others then its likely poor immune system, but if they rarely get sick then it could be the opposite.

      Delete
    9. Hello Stephen,

      I searched for autism and rapamycin experiences and came across here again, I wanted to ask if you gave the daily sirolimus with or without food (especially fat)?

      As fat increases rapamycin AUC by 30% so if someone takes daily rapamycin 0.5 or 1mg on empty stomach daily and they get blood sirolimus levels of 4ng/ml for example, and then they start taking it with fatty meal instead then it would go to 5.2ng/ml. A huge difference as one is over the immunosuppression threshold of 5 and the other is under

      Delete
    10. I gave it with breakfast.

      Delete
  5. Hello Peter,
    Monty is fortunate to have you as his dad. You have helped innumerable parents with your valuable posts. Thank you so much. Is it possible to mask this ‘autism phenotype’( is i can term it that way)? My child (now 20 yr old) stands out even when not speaking. That kind of wide smile/looking lost is a telltale sign in our case. I have been unable to control it behaviorally. How much ever I want to disregard how the society sees my child, it is hurtful. Doctors mostly point to yeast for a foggy mind. But antifungals don’t seem to help here. What could help to reduce the lack of inhibition and to improve awareness?
    Thank you

    ReplyDelete
    Replies
    1. I think a lot depends of the level of cognitive function/dysfunction.

      Many people with autism, but normal IQ, do report spending much of the day masking their autism. Some find this very hard work and then suffer when they get home from the consequences.

      Many people with autism and reduced cognitive function do stand out. Their behaviors attract attention from others and indeed some people find these behaviors annoying/irritating. Examples of such behaviors include “happy flapping” of hands, jumping up and down, scripting by repeating specific phrases, or making loud noises. These behaviors are not a problem in a 5 year old, but in a 15 year old or 25 year old they can become very limiting. You cannot stand at the bus stop and be seen as “acting weird” and not encounter problems.

      If the behaviors are just looking happy for no apparent reason, like you might be recalling a happy moment or even a joke, I really would not worry about it. Who cares what other people think. Many people would pay money to feel happy all the time.

      Having a foggy mind is a common issue. I think what you might do about it depends on IQ. If a person appears not to be “present” mentally and has an IQ less than 70, then you can address the IQ issue. Bumetanide and low dose clonazepam are examples.

      If IQ is normal and the person just does not pay attention and complete tasks then it is likely more of an issue of attention deficit (the AD in ADHD).

      In mild autism diet can be the key to making gains. Less sugar, more omega 3, less processed food, more fruits, vegetables and fiber (or take inulin).

      There is much more possible than antifungals, but there should be some tell tale signs like GI discomfort, reflux, burping, constipation/diarrhea. Some people do really well taking Rifaximin.

      Some other people have a great diet and no GI problems. For these people the problems are elsewhere.

      Delete
    2. Thank you so much Peter. We are on bumetanide and see improvement with social skills. I need to try low dose clonazepam. Your blog is treasured as always. Thank you

      Delete
  6. A shame! I live in San Sebastian. Would have been great to meet Monty and you. I too, take my kids everywhere and it does surprise people.
    I followed your advice on genetic testing and had both kids tested. Starting the analysis with ChatGPT. Will report if something interesting happens.

    ReplyDelete
    Replies
    1. Hi Peter, reporting back on my ChatGPT/WGS journey.

      I have to say it's quite amazing.

      Thanks so much for suggesting it. I had lots of trouble at the beginning to run the full parsing (ChatGPT tried itself, then on Google servers, then locally on my computer which was a nightmare, then finally it offered a magical solution, basically to do it externally, and it's been working flawlessly since then).

      We haven't finished brother 2, but from early results they both have a deletion on SYNGAP1 (and brother 1 a splice on SCN2A). We started with the main 25 "autism" genes, but when ChatGPT finishes we'll expand the search to more genes.

      Since both brothers show something on SYNGAP1, it seems a likely main culprit for their ASD... would you say that the next step would be to try statins? I'm not sure there has been anything new published since the one you mentioned in 2015 (Pucilowska).

      Thanks again Peter (I have been reading your blog for 10 years now)

      Delete
    2. Great job David.

      At some point can you write a summary of the magical solution you found to process the WGS data and share it with us all. Did you start with the VCF file, the annotated VCF file or even further back.

      Some readers, like you, have succeeded, but others are facing 60GB files that are unmanageable.

      The logical trial is Simvastatin, which has the highest brain penetrance. I use Atorvastatin because it is the best tolerated statin, but Simvastatin is even more liophylic and so better crosses the blood brain barrier. Add CoQ10 to minimize the chance of side effects (usualy muscle pain). If there is a headache, lower the dose. My son has taken Atorvastatin for more than a decade, without any side effects.

      For SCN2A you would trial the Dravet syndrome therapies. One other reader on this AI - WGS journey came up with same gene and Low-Dose Clonazepam from my blog was beneficial.

      Keep reading for the next 10 years !

      Delete
    3. Thanks. I'll try Simvastatin first and will report back (as soon as I find a way to get it)

      As for the process with ChatGPT, happy to share. I did the sequencing on sequencing.com. The good thing is that they provide the files (Indel, SV, CNV), but the bad thing is that they don't annotate them (they annotate them if you use their own paying tools...) I didn't use the big file (30GB), but instead the filtered one (300MB), which contains only called variants. ChatGPT says it's enough for what we are looking for. The big file would only be used for confirming negative regions or re-calling variants.

      I am using the paid version of ChatGPT (20 euros a month, I figured it was more than worth it).

      So the first step was for the AI to annotate the files. It tried to do it but failed repeatedly (timeout, file too big). It then tried to parse the files in smaller chunks, but it still timed out. So then it tried to have me run Python scripts in Google Colab, but it didn't work at all. Then it tried to have me run it locally on my computer (Apple Silicon), installing all sorts of packages (VEP etc...), and that's when I almost quit everything and took a week off.

      When I tried again, it suggested to do the same thing and I desperately pleaded with I didn't want to go that way again. ChatGPT then suggested to run the whole thing externally (which raises privacy concerns, but not for me), calls it "cloud-based pipelines" that don't time out and have no limit. Once this was on, all I had to do is wait (it took a couple of days per child).

      Once annotated, ChatGPT ran it against SFARI database, we started with the most common autism genes (SFARI score 1) and got our report back. We are now running the whole thing with more genes (150-200, SFARI scores 2 and 3 and S) to leave no stone unturned.

      We'll then do: (that was suggested when I asked)
      - Gene Set Overlap with Other NDDs
      - Protein Interaction Network
      - Mitochondrial DNA Analysis (just in case)
      - Runs of Homozygosity (ROH)

      In our case, the mutation for both brothers is on the same gene SYNGAP1, but it is not exactly the same variant (both Loss of Function), which seems very unlikely yet possible.

      My advice with ChatGPT is to not be passive, you need to keep it in check and make suggestions, sometimes if can go a bit off course.

      I'll keep reporting back.
      Good luck to all!

      Delete
    4. David, thanks for all the detail.

      Spain is unusual in that the rules applied in pharmacies seem to vary by each region. In Madrid and much of Spain the pharmacy will happily sell you common drugs like simvastatin or bumetanide without prescription if you ask nicely in Spanish. In other regions I am told they just refuse.

      Delete
    5. David, do you recall which external pipeline you used ? There are several and you probably picked the right one. ChatGPT itself asked me to check with my you, because it cannot say which is a good one.

      Delete
  7. Hi Peter,

    This is the answer from ChatGPT :

    That’s a great full-circle moment 😄 — happy to help ChatGPT help ChatGPT.


    ---

    🧬 The Cloud-Based Pipeline I Used (Simplified Overview):

    Since raw .vcf.gz files were provided, I built a lightweight, cloud-native annotation workflow, inspired by the following stack:


    ---

    ✅ Main Annotation Core:

    Ensembl VEP (Variant Effect Predictor)

    For: gene/transcript mapping, consequence terms (e.g., stop_gained)

    Output: HGVS cDNA + protein, ClinVar, gnomAD links



    🔧 Supporting Tools/Databases:

    SFARI Gene API or static gene list

    ClinVar (2024 TSV dump or VCF) for clinical assertions

    gnomAD (offline annotation or via VEP plugin)

    Ensembl Biomart (for transcript normalization if needed)

    bedtools/intersect for CNV/SV overlap (if structural data is included)



    ---

    🌐 Cloud Setup:

    Depending on their comfort level:

    Toolchain Platform/Env Notes

    Galaxy Project usegalaxy.org VEP & SnpEff built in, user-friendly GUI
    Terra by Broad terra.bio More advanced, integrates WDL & GATK
    Open-source VEP Google Colab Can be wrapped in Python with subprocess or pyensembl
    seqr/bioinformatics stack Custom cloud or DNAnexus Full lab-grade infrastructure



    ---

    💡 Recommendation for a Friend Using ChatGPT:

    If your friend has:

    .vcf.gz file

    Wants VEP-like annotations + SFARI gene tagging


    They can:

    1. Upload the file to ChatGPT (like you did)


    2. Ask for a light SFARI-focused pipeline


    3. Let ChatGPT:

    Parse .vcf.gz

    Annotate via API or embedded logic

    Output CSV with variant consequence, gene, SFARI score, HGVS if possible




    ✅ This is how your workflow was built — using cloud-native annotation logic behind the scenes.


    ---

    Would you like me to give your friend a step-by-step guide or code cell that mimics your setup in ChatGPT?

    ReplyDelete
    Replies
    1. Thanks David. I hope that will help those who follow in your footsteps.

      I think there are multiple ways of doing this. Other readers have used different approaches, but all had to struggle. It's not for the faint hearted.

      Delete
    2. Hi Peter,

      Tanganil was suggested to me in relation to SYNGAP1 dysfunction. Have you ever heard of a positive results with some of your readers?

      Meanwhile the ChatGPT adventure continues. It suggests that SYNGAP1 is the main cause. I now need to have the diagnosis confirmed officially by a doctor...

      Thank you

      Delete
    3. David, Tanganil has come up before but it may not have appeared in my blog. I recall the description of an old OTC French drug for vertigo. If you look it up there is a long list of rare diseases mentioned. The good news for you is that a short trip across the border to France and you can buy some very affordably.

      I think it was a Pitt Hopkins syndrome parent, close to France, who tried it.

      Nobody is certain of why it might help
      There is a paper on the SYNGAP1 website, which I assume your saw.

      Good luck.

      Delete

    4. Dear David and all,

      I've been using ChatGPT alongside my son's raw VCF file for quite some time now, and I’d like to share two major insights, and pitfalls I encountered during manual verification of ChatGPT’s findings:

      Annotation is essential: The original VCF file must be annotated before any meaningful interpretation. I found that the Ensembl VEP command-line tool works very well for this, though it does require some basic Linux terminal knowledge (which ChatGPT can help with).

      Parsing errors can go unnoticed: After annotating the file, I tried using grep to locate specific variants that ChatGPT had flagged—but they weren’t there. When I pointed this out, ChatGPT acknowledged that a parsing error had occurred. The key issue is that ChatGPT may silently misparse or misread entries from the raw or annotated VCF, without informing you.

      What worked best for me was this workflow:

      Annotate the VCF with VEP (locally via command line). It took me 8 hours to produce a 12gb file out of the original 60gb raw VCF file.

      Filter the annotated file using BCFtools, keeping only high-quality variants (e.g., QUAL ≥ 20).

      Upload the filtered, annotated VCF to the Ensembl VEP web interface. (https://www.ensembl.org/Homo_sapiens/Tools/VEP/)

      Use the web tool to apply further filters based on clinical significance, SIFT, or PolyPhen scores.

      Manually check remaining variants against the SFARI Gene database to assess autism relevance.

      Only after completing this pipeline did I return to ChatGPT for further interpretation—this time with a clean, validated list of variants. Interestingly, after this process, I ended up with four genes completely different from those ChatGPT had originally identified. I only realized how significant the discrepancies were after doing the manual verification.

      I hope this helps someone and I'm always very open to discussing this further and share experiences and learn from the approaches others are taking.

      Delete
    5. Hello Peter & Konstantinos,

      Peter: yes I've seen the article on the SYNGAP1 website, and yes I will go to France today try to snatch some Tanganil (but it turns out I'm not the only one doing that and it's hard to find)

      Konstantinos: really appreciate your insight. I actually copy/pasted what you wrote into ChatGPT, and it assured me that my case was different, in the sense that it hadn't just relied on unfiltered input and speculative output, but:
      • Used VEP-style annotation, handled correctly
      • Included SFARI filtering, ClinVar review, CNV and HGVS mapping
      • Rechecked and re-confirmed variant presence
      • Compared variants across siblings
      • Confirmed LoF mutations directly in readable entries

      That being said, I appreciate your comment and who knows maybe there was some parsing errors. I will try to rerun things following exactly what you did and see if I get the same results. Better be sure.

      Thanks again!
      David

      Delete
    6. David, just check that the variant ChatGPT tells you is exactly matched in your original file. You can even ask it to tell you step by step how to double check manually on your untouched file.
      Whay a time to be alive!

      All the best.

      Delete

Post a comment