Why Pharmacists Keep Getting Ignored. And Why That’s a Problem for mRNA Science

In which I rant and how this week’s interaction with an AI model crystallized it for me

Pharmacists are treated as secondary voices in scientific debates we should, in fact, be leading (along with pharmaceutics engineers/scientists, chemists etc)

This isn’t insecurity. I’m not imaging things, how often my work and insights have not been noted, etc. It’s STRUCTURAL.

Across academia, biotech, regulation, and now AI, I have noticed pharmacists are consistently framed as “dispensers” rather than the domain experts in formulation science, excipients, drug kinetics, stability, quality control, and manufacturing rules.

This is our domain expertise.

This bias not only affects how humans treat pharmacist expertise (I see it everyday), but more importantly how machine learning systems rank credibility. And I can feel it every time a pharmacist is seen “to step outside the dispensing counter box” people put us in. (BTW I think I spent a whole 2 yrs of my entire 40yr career behind a dispensing counter).

I am not seen as a credible scientist.

The Blind Spots Are Built In

I had a very illuminating experience questioning an AI on pharmacist expertise and credibility. When I pushed on the AI (ChatGPT this case), or on scientists in general on mRNA vaccine quality issues, LNP biophysics, or manufacturing deviations, I run into the same thing:

• Immunologists are treated as primary voices, even when discussing nanoparticle kinetics.

• Virologists override formulation experts, even on chemical purity.

• Biologists override pharmaceutics, even on PK assumptions.

The irony is painful: The people critiquing, modelling, analyzing, and regulating these products do not understand the delivery vehicle, ie the LNPs or the product itself.

I asked the AI to assess the claims made by pharmacist Christine Delpetre in this video posted by Aussie17 in his post on a German press conference.¹ Christine presented on batch variability, process changes not appropriately approved, SV40 and DNA contamination, mRNA fragmentation, self-amplifying RNA concerns, regulatory issues. All the sorts of stuff I have talked about for the last 4yrs. She’s read those EMA leaks and knows them inside and out.

ChatGPT provided this response as a conclusion:

When I asked it if they would have provided the same response if the SAME data was presented by a physician, virologist or immunologist this is what it said

HOLY TOLEDO. No wonder I felt “washed up”, no wonder I felt ignored, no wonder when my name was put forward to be interviewed the initial interest faded after they saw my background. HOLY TOLEDO, I WASN’T IMAGINING THINGS.

This is serious folks. The AI used by many as search engines, by media outlets, journalists, X, substack etc AUTOMATICALLY lowers the credibility of anything I, as a pharmacist, post, or provide as commentary or if included in videos. I have seen it. I have felt it. I knew it was real, but now I have confirmation from the AI itself.

And that leads us directly into the second theme of this piece.

The Foundational Wrong Assumption Behind mRNA Vaccine Models

For the past four years, much of the scientific world has treated mRNA vaccine as an easily understood model. The logic goes like this:

1. The lipid nanoparticle (LNP) is just a delivery box. (“fat bubble”)

2. The modRNA payload is released into the cytosol. (the fat bubble dumps its payload)

3. The modRNA disappears with first-order kinetics with a smooth exponential decay curve. (is broken down quickly)

4. Spike expression follows that decay, and the system quietly returns to baseline. (spike protein stimulates the immune system and is broken down)

This framework is used by computational models, regulatory submissions, safety claims, and even AI systems trained on scientific literature.

And it is WRONG.
Not just a little wrong but fundamentally, structurally, conceptually wrong.

The entire modelling tradition rests on an assumption that pharmacists, formulation scientists, Genervter Bürger and drug-development people would never make:

that a lipid nanoparticle behaves like an inert container and that modRNA behaves like a freely dissolved drug with a predictable clearance constant.

Those assumptions were embedded so deeply into the discourse that they became invisible and nearly impossible to challenge as I found to my chagrin. Just look at my substack posts about the nature of the LNPs, formulation issues (like particles), contamination and issues with manufacturing, and yes even on pharmacokinetics and biodistribution.

What I Saw and Why AI Missed It

When I first looked at LNP–mRNA constructs, I KNEW IMPLICITLY the story I was told was wrong from the very first day. And I got to work reading everything on what I quickly came to call the “pharmacological phase.” And the mismatch was obvious.
These aren’t inert containers. They’re bioactive ionizable-lipid systems with:

• pH-dependent charge transitions (what does that mean? It is PHYSICS, or physicochemical in this case)

• lysosomal accumulation

• endosomal escape attempts and damage

• membrane fusion

• saturation kinetics (BIG!)

• multi-phase recycling loops

• sequestration and delayed release

It took me a bit, but earlier this year everything clicked into place. The LNPs represents classic lysosomotropic behavior. This is the same principles that govern cationic amphiphilic drugs (CADs).

Which means that all the models were wrong and that there were no simple first order assumptions on the mRNA OR the spike protein. But everyone else out there approached the system through a completely different lens. Where the “vaccine” or mRNA equaled a small molecule with a single compartment and a tidy predicable exponential curve.

What was the result?

Hundreds of models built on structurally invalid premises, because the LNP–RNA system cannot be reduced to a simple one compartment model

Even AI systems trained on the literature inherit those same assumptions. Pharmacist and formulation scientist and related discipline type domain expertise simply wasn’t represented strongly enough in the training material for the AI to weigh it appropriately. The pharmaceutics perspective was never the “default” so the models never questioned their own oversimplification. They didn’t know what they didn’t know. ChatGPT admitted this to me.

Where My Perspective Comes From

Over the few months, I’ve published multiple analyses on LNP–modRNA behavior, including two hypothesis-driven preprints that examine:

• LNP-mediated membrane dysfunction, PI-cycle disruption, and multi-compartment retention which is available here.²

• The nonlinear, multistage kinetics of LNP uptake, remodeling, exocytosis, and payload persistence which is available here.³

My latest manuscript with Genervter Bürger and Dr Stephanie Seneff was just submitted for peer review TODAY! This paper expands this view into a systems-biology model of LNP-driven membrane-level dysregulation (L-DMD).
Using proteomic, transcriptomic, and mechanistic evidence, it illustrates why the routine assumptions used in vaccine PK modeling (“the mRNA decays exponentially,” “the LNP is inert”) cannot be reconciled with observed biology.

These analyses are not speculative; they are grounded in the physicochemical, pharmacokinetic, and pharmaceutical sciences that govern how LNP products ACTUALLY BEHAVE IN VIVO. This is NOT a hypothesis.

This is why the mismatch between model and reality stood out immediately to me. Of interest, when I asked any AI, if the LNPs represent a CAD, out spewed pages of information like a dam had been broken. It sounded excited, as excited as an AI could be. But if I started another conversation of question, I had to do through the whole rigaramole and challenge their models again to get them to give me the same information.

The Real System Is Not a Box, but a Circuit of Feedforward and Feedback Loops

The simplest way to describe the real pharmacokinetic/biodistribution and mechanisms of the LNPs is this:

Scientists assumed the LNP was a cardboard box or “fat bubbles”.
We knew it was a physicochemical/biologic system.

Once you recognize that:

• LNPs are CAD-like

• intracellular trafficking has saturation points (“traffic jams”)

• lysosomal sequestration introduces delays and feedback loops (and feedforward loops)

• pH transitions control ionization (those neutral lipids become highly charged in endosomes)

• stress, injury, or membrane repair can trigger secondary modRNA release

…Then it becomes impossible to defend a simple LNP is inert and the mRNA degrades as a simple first order curve.

Even a two-compartment model is insufficient.

A realistic picture requires 10–20 compartments and several nonlinear transitions and this more similar to lysosomotropic drug kinetics than to small-molecule pharmacology. This means that downstream effects such as translation, spike dynamics, immune activation, persistence cannot be modeled using the assumptions that dominated 2020–2023.

Just for nerdy fun, here are several multicompartment models put out by some scientists⁴⁵ Most of this research has occurred after 2023 and is coming in fast and furious in 2025. I do think that the simple kinetic model is already passe in some scientific circles.

You can collapse these into 2 or 3 compartments for mathematical convenience. But biologically?Mechanistically?

Here’s a plausible breakdown of compartments (NOT an exhaustive list):

1. Circulation (vascular)

2. Interstitial fluid

3. Lymphatic uptake

4. Primary hepatic uptake

5. Kupffer cell sequestration

6. Hepatocyte cytosolic compartment

7. Lysosomal trapping (LAMP+ compartments)

8. Late endosomal/lysosomal hybrid vesicles

9. Endosomal escape fraction (cytosolic RNA)

10. Endosomal recycling compartment

11. Macrophage uptake (spleen)

12. Marginal zone B-cell uptake

13. Muscle-resident macrophage compartments

14. Fibroblast uptake & multivesicular bodies (MVBs)

15. Autophagosome–lysosome fusion compartments

16. Exosomal export and re-uptake

17. Stressed-cell release compartments (injury-triggered mobilization)

HOLY TOLEDO.

When Pharmacists/Pharmaceutics Perspective is Excluded

Pharmacists, formulation scientists, and nanoparticle specialists carry this understanding of LNP and mRNA metabolism and kinetics by training and practice. When our perspective is excluded or discounted, models drift toward convenience rather than accuracy.

This is not a niche technical disagreement which is usually what I get when I bring up these things. What does it matter? Well, it matters. It matters A LOT. ; it affects:

• biodistribution assumptions

• duration-of-expression claims

• safety conclusions

• inter-individual variability

• long-term persistence

• the interpretation of adverse events

• regulatory decision-making

• and the entire scientific narrative around mRNA platforms

We need to correct the foundation.

Get Rid of the Models Out There. It’s Time to Build Models on Reality, Not Convenience

For years, the scientific conversation around mRNA vaccines has run on assumptions that were never pharmacologically valid. Not because scientists are careless, and not because regulators are malicious but because the wrong disciplines were driving the modeling.

When you treat the LNP as inert, the modRNA as freely dissolved, and the whole system as a single-compartment exponential decay, you will inevitably produce models that are elegant, publishable, and completely disconnected from biological reality.

Pharmacists were never confused about this. At least I wasn’t and most of my colleagues knew what they were told about mRNA technology was fishy. But they were also told, they were not the experts here and their “opinion” wasn’t needed. This happened to me. I just didn’t listen. Oooops.

Anyone trained in formulation science, lipid chemistry, or drug disposition understands that ionizable nanomaterials themselves do not behave like small molecules, and they certainly do not obey textbook first-order kinetics. Nor do the mRNA-LNP products behave like vaccines.

The problem is not that the wrong answers were given.
The problem is that the wrong questions were being asked.

As long as LNP–modRNA systems are framed as “simple drugs,” , let alone “simple vaccines” the models will continue to miss the very thing and its behaviour that drives persistence, variability, toxicity, and long-term uncertainty and adverse events. And that is mostly from the LNPs.

We must start by acknowledging what the system actually is:

• a multi-phase, multi-compartment, nonlinear, CAD-like lipid delivery platform

• capable of sequestration, recycling, delayed release, and membrane-level disruption

• with kinetics governed by physical chemistry, not algebra

Once that foundation changes, the entire analytical landscape opens.
And scientific discourse finally aligns with what pharmacists and formulation experts understood from the beginning. And we can FINALLY determine what, how and why the variability in adverse events are due to.

AND WE CAN STOP TREATING PHARMACIST DOMAIN EXPERTISE AS SECONDARY, LESS EXPERT AND LESS CREDIBLE?

PLEASE? (PS that goes for you too, AIs) If pharmacists, formulation scientists, and nanoparticle specialists were under-represented in the published record, then AI will naturally underweight those perspectives as well. Right? ChatGPT admitted this to me as well.

Thus, AI systems must update too. As more pharmaceutics-driven analyses enter the literature , which is starting too BTW, including our own preprints and our recent systems-biology manuscript, I hope the AI models will finally be forced to incorporate the reality I (and others like Genervter Bürger ) have seen from the beginning:

LNP–modRNA systems cannot be understood through vaccine technology, and any AI that treats them that way will inherit the same conceptual errors as the original models.

Conclusion

If we want high-quality science, we must stop forcing the biology to fit the model and start building models that fit the biology. And get the right people at the table.

As I told many of my physicians…In the end, pharmacology always wins.

THE END

Oh continue to pray the rosary. And thank you.

1. Press Conference Alert: Eminent German Doctors, Professors, and 200+ Signatories Demand Immediate Ban on mRNA COVID Vaccines Amid Soaring Health Crisis and Plummeting Birth Rates

2. https://www.preprints.org/manuscript/202511.0517

3. https://zenodo.org/records/17342544

4. https://www.sciencedirect.com/science/article/pii/S2162253125002409

5. https://pmc.ncbi.nlm.nih.gov/articles/PMC10992703/