Frequency Therapeutics — Hearing Loss Regeneration

[FGG]

I think age-related hearing loss trial should have no problem passing but I'm a bit worried about the severe hearing loss trial. I still believe FX-322 will do something for severe hearing loss sufferers e.g. the ability to hear high frequencies a lot better.

Why do you think the age-related trial will go better than the severe arm? I think the complete opposite.

The severe arm is likely to have much more IHC damage and should be better able to show hearing improvements with the drug.

The age-related arm could be another wash imo.

 

[kiki]

Wonder how much that would translate to in decibels.

I think so, too. How much is the physical displacement converted to dB?
What kind of calculation formula would it be?

 

[FGG]

Nice find, I will definitely watch it soon.

Regarding the auditory nerve, does anyone here know if strands of the nerve are frequency specific when electrical impulses are en route to the brain?

Or is it more like an electric wire where as long as there is continuity, a signal will get through?

Pardon my ignorance.

From my understanding, in a normal ear, the impulses are generated from the IHCs and are not changed as they go through the auditory nerve.

Demyelinating disorders (e.g. MS, GBS) can change the impulses somewhat from that point during travel through the nerve but that's not in a "frequency specific" kind of way.

 

[kiki]

Ok here's my theory + diagrams...

Let's set a few parameters first:

The makeup of the IHC + OHC in the cochlea:
A single human cochlea generally has about 3500 ROWS consisting of 1 IHC + 3 OHC. This enables most people to hear from 20 Hz - 20 kHz. So, back of the napkin math tells us that each row is responsible for roughly 5.7 Hz sensitivity. Since the biology cochlea isn't based on the Hz measurement, it stands to reason that the metric itself doesn't matter all that much in figuring out what each row can actually detect. Especially since the human cochlea is able to distinguish between two frequencies 1 Hz apart.

What I have come to understand is that each row has a lot of overlap in terms of sensitivity to frequencies. It appears that there is considerably more overlap in sound frequency sensitivity amongst IHC, but not as much in terms of decibels; they provide a flat signal to the brain. The OHC do not have the same level of overlap in frequencies, and are much more fine tuned to specific frequencies; due to their mechanical nature, they're able to tune based on the power (decibels) of sound received by the cochlea. When a sound is faint, they can turn a signal "up", when it is loud, turn it down. Etc.

Diesel's diagrams:

For the sake of example, let's assume that each row of hair cells translates to roughly 5 Hz of frequency sensitivity. In the first example, we see a sample of healthy cochlea, centered on the 4 kHz location +/1 20 Hz.

The far left figure indicates the frequency that the IHC can still "detect", and since they are green, the IHC row is fully sensitive to the entire sample. The IHC row indicates that the cell is present. Green = present. Red = dead/destroyed. The next three OHC rows indicate the adjacent OHC. Again, all are green and present. The far right row indicates the sensitivity of the OHC SET (all three) at the specified frequency.

View attachment 44429

Ok, let's look at what happens when damage occurs above and below the 4000 Hz row. Leaving the 4000 Hz cells alive and healthy:

IHC: The 4000 Hz IHC is still able to detect adjacent frequencies above and below 4000 Hz to an extent. At 3985 Hz + 4015 Hz, sensitivity is partial.

OHC: The OHC cannot detect beyond its missing neighbor.

In this example, the audiogram would look NORMAL at 4000 Hz. Word score may be ok as well.

View attachment 44430

Let's look at another example:

Here, the 4000 Hz row stays alive, as does the 4020 Hz row.

IHC: Due to the larger range of sensitivity of the IHC, they are able to continue to detect sound well from 3990 Hz - 4020 Hz and beyond. YET, there are IHC missing from 3 frequencies in the middle.

OHC: Again, lacking range in sensitivity, an audiogram test at 4005 Hz - 4015 Hz may reveal a deficit.

View attachment 44434

So, let's take the same example and pick off an OHC at 4000Hz:

IHC: Coverage still looks good. Word Score should be retained.

OHC: May be an issue apparent on the audiogram at 4000 Hz.

View attachment 44435

Let's mix it up. What if only Rows at 3980 Hz and 4020 Hz survive?

IHC: Severe deficit at 4 kHz. But, overall coverage in this sample is still not terrible.

OHC: Severe deficit at 4 kHz. Overall deficit from 3985 Hz - 4015 Hz. Would definitely appear on audiogram.

View attachment 44436

Ok, one last sample re: damage to bring it home.

Keep the outer rows normal. Damage the 4 kHz OHC only.

IHC: Almost total coverage. WR score should remain excellent.

OHC: Obvious deficits. Would be visible on audiogram. I believe this would show a mild loss if the research correlates from the "Diesel Rat" study.

View attachment 44437

If there is only one OHC, will the right end be red?

 

[Lucifer]

Why do you think the age-related trial will go better than the severe arm? I think the complete opposite.

The severe arm is likely to have much more IHC damage and should be better able to show hearing improvements with the drug.

The age-related arm could be another wash imo.

In the age-related hearing loss trial they are testing mild-moderately severe age-related hearing loss. I think it should do better as most of the damage is at the high-mid frequencies whereas in the severe hearing loss trial the damage is more likely to be across all the frequencies. Also the severe hearing loss trial participants may have little to no support cells compared to the age-related hearing loss trial.

 

[FGG]

In the age-related hearing loss trial they are testing mild-moderately severe age-related hearing loss. I think it should do better as most of the damage is at the high-mid frequencies whereas in the severe hearing loss trial the damage is more likely to be across all the frequencies. Also the severe hearing loss trial participants may have little to no support cells compared to the age-related hearing loss trial.

Frequency Therapeutics needs to demonstrate IHC regeneration if they are continuing with the single arm though...

And support cell loss affects more of the profound group which are not part of the severe arm.

 

[Lucifer]

Frequency Therapeutics needs to demonstrate IHC regeneration if they are continuing with the single arm though...

And support cell loss affects more of the profound group which are not part of the severe arm.

I see where you're coming from. I agree with your first point and a bit of your second point but with the severe arm if the damage is across all frequencies that means FX-322 needs to do more work at regrowing IHCs in all frequencies compared to mild-moderately severe arm where most of the damage is at the high-mid frequencies.

I'm just worried that for FX-322 to be more effective for the severe hearing loss trial they would need to come up with a new delivery method to reach deeper in the round window. With the current delivery method the mild-moderately severe hearing loss should still be fine.

I do hope the severe hearing loss trial results look very good even with the current delivery method.

 

[FGG]

I see where you're coming from. I agree with your first point and a bit of your second point but with the severe arm if the damage is across all frequencies that means FX-322 needs to do more work at regrowing IHCs in all frequencies compared to mild-moderately severe arm where most of the damage is at the high-mid frequencies.

I'm just worried that for FX-322 to be more effective for the severe hearing loss trial they would need to come up with a new delivery method to reach deeper in the round window. With the current delivery method the mild-moderately severe hearing loss should still be fine.

I do hope the severe hearing loss trial results look very good even with the current delivery method.

The age-related arm isn't really a "mild to moderate arm" since the inclusion parameters were literally anything from mild to severe loss (up to 4000 Hz). Most age-related loss is much greater and the higher frequencies, though (could even be profound in the EHF in many of the oldest patients which could further complicate things). I think the group data will be very murky even if some of the individual data could be interesting/useful.

The reason the severe arm holds much more promise imo is that, as a group, you are much more likely to get statistically significant word score changes even if there are a few "odd men out" who have worse loss at lower frequencies.

 

[Lucifer]

The age-related arm isn't really a "mild to moderate arm" since the inclusion parameters were literally anything from mild to severe loss (up to 4000 Hz). Most age-related loss is much greater and the higher frequencies, though (could even be profound in the EHF in many of the oldest patients which could further complicate things). I think the group data will be very murky even if some of the individual data could be interesting/useful.

The reason the severe arm holds much more promise imo is that, as a group, you are much more likely to get statistically significant word score changes even if there are a few "odd men out" who have worse loss at lower frequencies.

I agree that the severe hearing loss trial will be more consistent as a group as most should be getting similar results as the others whereas the age-related hearing loss will have more inconsistencies.

It will easier to rule out whether FX-322 works or not in the severe hearing loss trial.

 

[kiki]

In this thread, I think "IHC-preference" is a common recognition.

Is it because the process of regeneration from progenitor cells repeats the process of development in the womb?

Isn't there a possibility of OHC regeneration even with multiple injections at appropriate intervals?

Was the 10 dB or 5 dB improvement of Phase 1/2 an error? (If that anecdote was true that a participant in Phase 2a improved by 10 dB in 3 bands, would that also be an error?)

 

[Zugzug]

In this thread, I think "IHC-preference" is a common recognition.

Is it because the process of regeneration from progenitor cells repeats the process of development in the womb?

Isn't there a possibility of OHC regeneration even with multiple injections at appropriate intervals?

Was the 10 dB or 5 dB improvement of Phase 1/2 an error? (If that anecdote was true that a participant in Phase 2a improved by 10 dB in 3 bands, would that also be an error?)

It's speculation. If the severe trial results in very successful WR results and only small audiogram changes, the theory will hold greater merit.

 

[Lucifer]

It's speculation. If the severe trial results in very successful WR results and only small audiogram changes, the theory will hold greater merit.

Even with small audiogram changes, multi-dosing may allow more OHC regrowth but they need to space out dosing to at least monthly.

 

[Zugzug]

Even with small audiogram changes, multi-dosing may allow more OHC regrowth but they need to space out dosing to at least monthly.

A new rule that I have for myself is to only focus on what's proven. Incremental speculation is fine based on the evidence, but everything becomes a low probability mess if we go too far down rabbit holes.

 

[MrCrybaby]

IMO even if this drug is only useful in severe cases of hearing loss it will be a success.

Once there is a product on the market there will be incentive for better testing tech to be developed and the cash flow can be put into refining this drug and developing other candidates (assuming that this whole PCA process works as well in the clinic as the lab).

 

[FGG]

A new rule that I have for myself is to only focus on what's proven. Incremental speculation is fine based on the evidence, but everything becomes a low probability mess if we go too far down rabbit holes.

I fully agree that you can't take speculation as fact without more data (e.g., the severe arm) but I don't think we can apply probabilities when we don't know the mechanism.

This is where math and science are not a perfect fit without knowing all the variables.

For my UTI and diabetic dogs example, for instance, just looking at the math you might conclude insulin is less effective in females, when actually female dogs just get more UTIs (with increases insulin resistance when active).

 

[serendipity1996]

IMO even if this drug is only useful in severe cases of hearing loss it will be a success.

Once there is a product on the market there will be incentive for better testing tech to be developed and the cash flow can be put into refining this drug and developing other candidates (assuming that this whole PCA process works as well in the clinic as the lab).

Agreed. I think the priority is to get a minimum viable product onto the market for a condition where there is a massive unmet need and no other drug options. Then it can be improved on further.

 

[Zugzug]

I fully agree that you can't take speculation as fact without more data (e.g., the severe arm) but I don't think we can apply probabilities when we don't know the mechanism.

This is where math and science are not a perfect fit without knowing all the variables.

For my UTI and diabetic dogs example, for instance, just looking at the math you might conclude insulin is less effective in females, when actually female dogs just get more UTIs (with increases insulin resistance when active).

Ahh, my eyes, they burn. There's a few things you never bring up to someone. Their family (negatively), finances, politics, and why their love for mathematics is a sham.

Seriously though, I would argue math is specifically designed to answer the UTI question. Mathematicians are deep investigators, interested in hard proofs of claims. In fact, one could argue that the biggest fault of mathematicians is that they are too interested in mechanisms to the point where they never graduate past them.

I would argue that a mathematician's job is to actually say "wait, that's a correlation, not a proof!"

I would also math is at the heart of science. Just as an easy example, it is quite common to perform Z-tests when comparing proportions of groups (treatment vs placebo) for large samples. All of the justification for this is based on Pierre-Simon Laplace's proof of the Central Limit Theorem (1810), which basically proves that samples averages behave like bell-shaped curves for large sample sizes.

Then there's the Student's t-test for small samples. Funny thing, the reason it's called "student" is actually because the statistician (William Sealy Gosset, 1908) who discovered the T-distribution (bell shaped curves with fatter tails for small sample sizes) was encouraged to hide his identity. This test is used everywhere. It was used a few times for FX-322 in Phase 1/2. It's highly useful because it's one of the few tests that is valid for small sample sizes without a "convergence" aspect going on like in the Z-test.

Finally, the point that math does not work without knowing all of the variables, this is precisely what the point of differential equations is. In most differential equations, one takes laws of physics along with spatial or time data conditions to solve for the unknown variables.

Quite literally, nothing in these clinical trials makes any sense at all without math or statistics lol.

I'm just playing around here, BTW. I know you're not a disrespectful person at all lol.

 

[FGG]

Ahh, my eyes, they burn. There's a few things you never bring up to someone. Their family (negatively), finances, politics, and why their love for mathematics is a sham.

Seriously though, I would argue math is specifically designed to answer the UTI question. Mathematicians are deep investigators, interested in hard proofs of claims. In fact, one could argue that the biggest fault of mathematicians is that they are too interested in mechanisms to the point where they never graduate past them.

I would argue that a mathematician's job is to actually say "wait, that's a correlation, not a proof!"

I would also math is at the heart of science. Just as an easy example, it is quite common to perform Z-tests when comparing proportions of groups (treatment vs placebo) for large samples. All of the justification for this is based on Pierre-Simon Laplace's proof of the Central Limit Theorem (1810), which basically proves that samples averages behave like bell-shaped curves for large sample sizes.

Then there's the Student's t-test for small samples. Funny thing, the reason it's called "student" is actually because the statistician (William Sealy Gosset, 1908) who discovered the T-distribution (bell shaped curves with fatter tails for small sample sizes) was encouraged to hide his identity. This test is used everywhere. It was used a few times for FX-322 in Phase 1/2. It's highly useful because it's one of the few tests that is valid for small sample sizes without a "convergence" aspect going on like in the Z-test.

Finally, the point that math does not work without knowing all of the variables, this is precisely what the point of differential equations is. In most differential equations, one takes laws of physics along with spatial or time data conditions to solve for the unknown variables.

Quite literally, nothing in these clinical trials makes any sense at all without math or statistics lol.

I'm just playing around here, BTW. I know you're not a disrespectful person at all lol.

I feel like I just told you that your baby is ugly (or not useful ) and I wasn't saying that at all. Of course the math is important and vital. You can't fully prove anything without it.

I guess a better way to express my point might be that it's hard to apply the math (in terms of percentages) without looking closely at the individuals (i.e. and figuring out what "exactly" makes someone a responder or a non-responder which is more a question of biology than math).

And the reason I brought up the UTI example specifically is because it's a very non obvious co-factor. You could spend years looking at hormone differences and body fat distribution before you ever considered the urinary tract. Math could tell you female gender was related but you can't answer the how or why with it.

The math is illuminating for sure and a big part of the analysis but the full picture of the anatomy and physiology is needed to use (the very valuable) tool correctly. And I know you weren't arguing otherwise I was just pushing back at your statement that multi factorial considerations make something "low probability" unless I misunderstood your point.

But please don't think I was saying math wasn't useful. It's clearly extremely useful.

 

[Zugzug]

I feel like I just told you that your baby is ugly (or not useful ) and I wasn't saying that at all. Of course the math is important and vital. You can't fully prove anything without it.

I guess a better way to express my point might be that it's hard to apply the math (in terms of percentages) without looking closely at the individuals (i.e. and figuring out what "exactly" makes someone a responder or a non-responder which is more a question of biology than math).

And the reason I brought up the UTI example specifically is because it's a very non obvious co-factor. You could spend years looking at hormone differences and body fat distribution before you ever considered the urinary tract. Math could tell you female gender was related but you can't answer the how or why with it.

The math is illuminating for sure and a big part of the analysis but the full picture of the anatomy and physiology is needed to use (the very valuable) tool correctly. And I know you weren't arguing otherwise I was just pushing back at your statement that multi factorial considerations make something "low probability" unless I misunderstood your point.

But please don't think I was saying math wasn't useful. It's clearly extremely useful.

Lol, yeah, I think I knew what you meant. My source of disagreement was not so much that a mathematician would figure everything out, but that their mindset would be "the numbers say women have evil bodies that hate insulin. That's a rap, QED." (If one did think in this way, it would probably be from some deep-rooted misogyny as opposed to proper technique.) No question, it would require someone with an understanding of the human body to investigate this issue.

Re: multi factorial: Let me try to explain my point better. Suppose we have proven facts. Then, of course, a necessary next step is to theorize through biological considerations what some new proven facts might be. If we assume the current proven facts and ask ourselves about the probability of a bunch of immediate next steps, this makes sense.

My point is that if a long "tree" of unproven events is constructed, the probability becomes very low. For example, if the severe trial shows massive WR score improvements and no audiogram changes, the next step is to prove that IHC are the main thing being targeted.

But say we got ahead of ourselves before proving this. Say we said, "well, IHC are at least the main target, but OHC could be involved. Then if OHC are involved, they will grow from multiple doses, well-spaced out. Then if that happens, major audiogram changes will in the >= 8 kHz range. Then after this occurs, we should be able to reduce the time between doses and push it even deeper. Then IHC in the 6-8 kHz will grow. Then on multiple passes, the OHC at 6-8 kHz will grow."

We still haven't proven that IHC are the main thing being targeted. If this makes sense, I'm suggesting that from a given point of knowledge, we consider branching out in multiple single branches, but resist the urge to pursue one branch chain with multiple branches on it. I'm sorry if that doesn't make any sense.

This is how I'm approaching it anyways. I've made this type of mistake many times in my own research over the years. I'll think something is true and instead of proving it with dumb, useless, boring cases, I'll want to prove the end goal and fail. Often times, in research math, a good strategy for proving something (or even heuristically learning if it's even true or not) is to start with stupidly easy examples and see if you can either prove the theorem or come up with a counter example. If you can, you move on to a slightly harder case.

How dare you call my baby ugly! BTW, my actual baby, my cat, I often joke is completely average looking.

 

[FGG]

Lol, yeah, I think I knew what you meant. My source of disagreement was not so much that a mathematician would figure everything out, but that their mindset would be "the numbers say women have evil bodies that hate insulin. That's a rap, QED." (If one did think in this way, it would probably be from some deep-rooted misogyny as opposed to proper technique.) No question, it would require someone with an understanding of the human body to investigate this issue.

Re: multi factorial: Let me try to explain my point better. Suppose we have proven facts. Then, of course, a necessary next step is to theorize through biological considerations what some new proven facts might be. If we assume the current proven facts and ask ourselves about the probability of a bunch of immediate next steps, this makes sense.

My point is that if a long "tree" of unproven events is constructed, the probability becomes very low. For example, if the severe trial shows massive WR score improvements and no audiogram changes, the next step is to prove that IHC are the main thing being targeted.

But say we got ahead of ourselves before proving this. Say we said, "well, IHC are at least the main target, but OHC could be involved. Then if OHC are involved, they will grow from multiple doses, well-spaced out. Then if that happens, major audiogram changes will in the >= 8 kHz range. Then after this occurs, we should be able to reduce the time between doses and push it even deeper. Then IHC in the 6-8 kHz will grow. Then on multiple passes, the OHC at 6-8 kHz will grow."

We still haven't proven that IHC are the main thing being targeted. If this makes sense, I'm suggesting that from a given point of knowledge, we consider branching out in multiple single branches, but resist the urge to pursue one branch chain with multiple branches on it. I'm sorry if that doesn't make any sense.

This is how I'm approaching it anyways. I've made this type of mistake many times in my own research over the years. I'll think something is true and instead of proving it with dumb, useless, boring cases, I'll want to prove the end goal and fail. Often times, in research math, a good strategy for proving something (or even heuristically learning if it's even true or not) is to start with stupidly easy examples and see if you can either prove the theorem or come up with a counter example. If you can, you move on to a slightly harder case.

How dare you call my baby ugly! BTW, my actual baby, my cat, I often joke is completely average looking.

Ah, so you were saying the probability of drawing a hard conclusion was low, not that you were speculating on the probability of a particular mechanism then? If so, that makes total sense and I agree.

Btw, the UTI/Diabetes story wasn't to imply that I thought mathematicians would draw ridiculous unproven conclusions (you guys aren't flippant like that!) from data but rather that they way the data is framed, influences how it is perceived. And statistics can be used in this way to confuse rather than clarify.

For example, if you had population stats on dog gender and insulin resistance you might have a harder time narrowing in on the issue than if you instead had stats on the presence or absence of a UTI and insulin resistance. And that difference in analysis depends on including as many known variables as possible. You always have to very carefully look at the responders and non responders and figure out why and not (imo) make any general drug claims until you do. Even then with the lack of good testing, it's tricky.

Your point is valid that the math only proves correlation but, in some cases, it can effect the direction and conclusions people draw from the data. But of course, you also need math to draw the conclusions.

Anyway, I do agree the OHC is speculation/looking at studies and talking out loud about possibilities. The IHC regrowth seems a given if you believe Phase 1 isn't fraud and, as they are moving on with the single dose protocol, it's likely we wouldn't know what it does for OHCs with various other protocols until it's released if the later trials reproducibly show this effect.

 

[Keith Handy]

BTW, my actual baby, my cat, I often joke is completely average looking.

"Look at you! What an average-looking kitty you are!"

 

[kiki]

Anyway, I do agree the OHC is speculation/looking at studies and talking out loud about possibilities. The IHC regrowth seems a given if you believe Phase 1 isn't fraud and, as they are moving on with the single dose protocol, it's likely we wouldn't know what it does for OHCs with various other protocols until it's released if the later trials reproducibly show this effect.

Thank you. There is still hope.

 

[Zugzug]

"Look at you! What an average-looking kitty you are!"

You tell me my cat doesn't look exactly like the cat on the food bag. If 1 is ugly, 10 is attractive, my cat is a perfect 5. But her level of vindictiveness, evilness, and deceptiveness are 10/10. In this picture, she knows she's in trouble with daddy because she's not allowed near the counters.

 

[FGG]

You tell me my cat doesn't look exactly like the cat on the food bag. If 1 is ugly, 10 is attractive, my cat is a perfect 5. But her level of vindictiveness, evilness, and deceptiveness are 10/10. In this picture, she knows she's in trouble with daddy because she's not allowed near the counters.

View attachment 44449

I blame the media and unrealistic cat expectations. Your cat is easily an 8. And I have seen a lot of cats!

 

[Zugzug]

Your point is valid that the math only proves correlation but, in some cases, it can effect the direction and conclusions people draw from the data. But of course, you also need math to draw the conclusions.

To be honest, I'm actually happy that a biologist is pointing this out for once, so I'll happily take the hit for mathematicians if it improves scientific conclusions. In my experience, often times it's the exact opposite. The mathematician proves something with theory and logic only. The statistician then relaxes things a bit to perform a real study. The conclusion is then no implication of a causative relationship. The biologist assumes a causative relationship. The mathematician gets victim-blamed (I'm joking, of course -- that's some hyperbolic language) and everyone concludes that math is stupid and pointless. Of course, the biologist thinks the mathematician is an idiot because they can't work with their hands and actually demonstrate real molecular reactions. which is fair. So I am not implying that the mathematician is strictly better than the biologist or something.

Really, from Phase 1/2, I'm not a victim. The only thing those p-values concluded was that if the null hypothesis was true, there would be a very small percentage of observing those second word scores. Nothing more, nothing less. Except, Zugzug, not math, is the reason why I speculated further about the drug. I'm sticking to what's proven: The 3 super responders were not normal without something going on. That's all I know for now.

BTW, even stats is not really proving claims from data. It's quite different from math, as it involves significance levels and judgement calls based on levels of uncertainty. There's also often judgement involved in deciding if statistical assumptions are met. Math is straight logic, built from axioms only.

 

[Keith Handy]

You tell me my cat doesn't look exactly like the cat on the food bag. If 1 is ugly, 10 is attractive, my cat is a perfect 5. But her level of vindictiveness, evilness, and deceptiveness are 10/10. In this picture, she knows she's in trouble with daddy because she's not allowed near the counters.

View attachment 44449

Our cat used to jump off the kitchen counter before we entered. We'd go to the kitchen, hear a "thud," and arrive just in time to see a perfectly innocent cat looking up at us as if saying, "I've been down here the whole time!"

 

[Diesel]

Let's get back on topic.

 

[FGG]

To be honest, I'm actually happy that a biologist is pointing this out for once, so I'll happily take the hit for mathematicians if it improves scientific conclusions. In my experience, often times it's the exact opposite. The mathematician proves something with theory and logic only. The statistician then relaxes things a bit to perform a real study. The conclusion is then no implication of a causative relationship. The biologist assumes a causative relationship. The mathematician gets victim-blamed (I'm joking, of course -- that's some hyperbolic language) and everyone concludes that math is stupid and pointless. Of course, the biologist thinks the mathematician is an idiot because they can't work with their hands and actually demonstrate real molecular reactions. which is fair. So I am not implying that the mathematician is strictly better than the biologist or something.

Really, from Phase 1/2, I'm not a victim. The only thing those p-values concluded was that if the null hypothesis was true, there would be a very small percentage of observing those second word scores. Nothing more, nothing less. Except, Zugzug, not math, is the reason why I speculated further about the drug. I'm sticking to what's proven: The 3 super responders were not normal without something going on. That's all I know for now.

BTW, even stats is not really proving claims from data. It's quite different from math, as it involves significance levels and judgement calls based on levels of uncertainty. There's also often judgement involved in deciding if statistical assumptions are met. Math is straight logic, built from axioms only.

I didn't look at it from that angle. That would be totally unfair mathematician blaming. I agree.

I actually think mathematicians and biologists (apart from their background and career choice even) seem to have different strengths/skill sets.

The mathematicians are much better at logical consistency and the biologists are better at trying to piece together data in novel or extremely high variable situations. It's a great team in a lot of ways (in fact you obviously need both) but they sometimes practically need a translator between the two disciplines.

I never claimed math was stupid or pointless though!

 

[Zugzug]

Let's get back on topic.

View attachment 44450

Patient A improved about as much as Patient B in absolute WR percentage, except patient A took the drug and patient B took the placebo. Patient A had a lot of room for growth, while Patient B had hardcore ceiling effect. That's bad news. Change my mind.

 

[Street Novelist]

Is Frequency Therapeutics still sticking with its timeline of the next trial starting early next year?

If so, they will have to start recruiting soon; maybe in June/July.