Need to read this asap, but 6.5% is an exceedingly small effect size and not robust given many confounding developmental and technical factors. The typical coefficient of error between volumes of even large brain regions in genetically identical mice is usually well over 5%. And I suspect the transgene was only tested on one genetic background—-usually C57BL/6J.
Will emend this post tomorrow with corrections after reading carefully. I can say with reasonsble assurance that no one has measured and weighed more mouse brains than I have ;-) (except you John Wong)
You should have read the paper (or at least the abstract) first, as they used macaque and human organoid models, too. :) But yes, the overall effects of the region of DNA in question (an enhancer) are fairly small.
I have read it now. Comment below. As to an increase in mouse neocortical volume—they have ZERO data. Look at their figures 1g and 1h and weep for the crap reviewing.
> can say with reasonsble assurance that no one has measured and weighed more mouse brains than I have
Out of curiosity how many would that be? Thousands? Tens of thousands? How long does it take to measure a mouse brain? When you do it does someone remove them for you or you also have to remove them before measuring?
I guess that whatever recognises the enhancers must be conserved ish in mice for us to be able to drop this region into the mouse genome. That might be interesting alone apart from the functions of the frizzleds. Always a curious question about how much new “machinery” we have in humans, vs us being a specific configuration of common machinery (i.e., if we were to swap out introns/regulatory regions from a mouse with human, and assuming we don’t screw up any checkpoints, how far would it get with looking/acting like a human?).
Just reviewed the paper. Extremely elegant flamingo dance. But let me reserve my comments to the Hacker News theme of an enlarged brain (or really an enlarged neocortex). The key panels on this topic are Fig. 1g and 1h. OMG, the dissections of these "best examples" are sad–– the olfactory bulbs are serious torn and paraflocculi are detached. The histogram that is Fig 1h has a grand total of 6 measurement for the human transgenic mouse line and 4 (four) points for the control. Hey when you have a statistically significant result why mess it up by evaluating 10 of each of the two groups. You might not get published in Nature or airplay on HN, and would that not be a shame?
As for the measurement of brain size, it is a planimetric projection of the area of the dorsal surface of the neocortex. This is a lame way to do morphometry. Squish the brain a bit and the area will expand beautifully—and by way more than 6.5%.
One last comment on the genetics of these animals. The are probably incorrectly stated to be at least F8 progeny of the mixed 129-B6N embryonic stem cells. I hope they mean N8 backcross progeny––that is to say, 8th generation congenic lines. But in this case they appear to have backcrossed weirdly to a different type of B6; the standard C57BL/6J strain. All of this means that even in the best case, they have three different genomes banging around in supposedly co-isogenic cases and their controls: 1. chunks of 129 strain chromosomes that will still be common even at 8 generations, 2. chunks of B6N chromosomes that will also be common at 8 generations, and of course the B6J background strain. You would have to carry out sparse whole genome sequencing or use the GigaMUGA array to unconfound the genetics in this study.
Highly linear in vivo and close to 1. Brain has a lot of fat but not a big deal in mice. Only fine point is that the density of brain varies depending on fixation protocol but almost always by less than 5% if you get the osmolarity of the solution just right.
For the claim of increased neocortical size—-the topic of this HN thread, they show 4 controls versus 6 transgenic mice. Not sure if they did this blinded. Would depend mainly on when they did the genotyping.
The question remains if the same genes command an increase of the braincase, and what potential issues (including limited growth) happen if it's not the case
Meta comment: 11/13 comments currently are pop-culture references (Flowers for Algernon, Pinky and the Brain, other movies) or generic phrases and only 2 comments relate to the actual article. I've been on HN for over a decade, and this is the worst it's ever been.
Comments still turn up useful insights that you don't usually catch on other places, like the top one. I think the trend of low quality AI related crap getting to top of the front page is more annoying.
Here's a formula you can try :
Title : "I made a AI powered insert tool use-case"
Article: "Hey look what I did ! By hooking up insert LLM model to this insert random tool in a few hours of vibe coding I got it to do something that vaguely resembles something useful and made no effort to evaluate performance/compare to anything. Here's my toy demo example."
Should be front page material, going on the last few months.
Downvote, flag, and move on. These people don't understand the difference between HN and Reddit and there's always going to be more influx than you have bandwidth to talk out of it.
> There are much better professional discussions on
That is the random peak that results from lucky combinations in some corners of a space - it is not owing to how the channel works, it is not owing to the culture it promotes.
Comparing the best of the best of Reddit against "most HN threads" isn't a good argument. Either compare the best of the best of both (I'd argue that HN comes out on top on most tech / programming related issues, and Reddit comes out on top more on more general topics), or compare averages to averages (Here HN comes out on top in my view).
The comparison of best-of-the best is not that useful in practice because you have to filter through so much noise to find it, so the average case is more useful in practice.
I am aware of the guidelines, considering how long I've been here. However, I have offered proof, that over 80% of the comments do not contribute anything and are exactly like those of Reddit.
Given the standard practice in lab experimentation using mice (which is to "dispose" of them after their usefulness for a test is done) I'm extremely troubled by the idea of gene editing experimentation using mice to make them more intelligent (their standard intelligence would already be surprising to many people).
While I already think it's unethical to use animals the way we do, the idea that we're experimenting with how close we can get to bestowing them with human intelligence before killing them is saddening.
That is neat but doesn't mean much by itself without some kind of test on cognition. The size of a creatures brain doesn't really correlate with intelligence very well.
So when you give the mouse some human DNA, do you then need to send the proposal to the Institutional Review Board, since these mice are a tiny bit human?
feel like if they were smarter they'd be far less scary. is not the main fear with sharks that they'll mistake you for something they actually want to eat?
comment on the comments
geneticly spliced animals and humand are now real, but for many years have occupied a specific place in sci fi, where there are enhanced animals and humans, hybrids, and we have a huge subculture of "furryies" so the headline is pushing lots of buttons. The thing is that it's going to happen, and the pieces are here right now to make hybrids, and even alter a life forms DNA after birth.
The point bieng if anybody is getting triggered by the whole kids and smart phones thing, best just think of that as quaint , as there is a massive market for some little adorable critter that can talk.....brace yourselves
and nothing can make that possibility go away now
(:shrug: I get it... Sore point... The apocalypse isn't a humorous thing to joke about, what with it being in the midst of actually happening and all ... but if I don't try to find myself a little giggle here and there once in a while I'ma literally go insane, so whatever. Down-vote away. Don't really care anymore.)
Need to read this asap, but 6.5% is an exceedingly small effect size and not robust given many confounding developmental and technical factors. The typical coefficient of error between volumes of even large brain regions in genetically identical mice is usually well over 5%. And I suspect the transgene was only tested on one genetic background—-usually C57BL/6J.
Will emend this post tomorrow with corrections after reading carefully. I can say with reasonsble assurance that no one has measured and weighed more mouse brains than I have ;-) (except you John Wong)
You should have read the paper (or at least the abstract) first, as they used macaque and human organoid models, too. :) But yes, the overall effects of the region of DNA in question (an enhancer) are fairly small.
The preprint is available here: https://www.biorxiv.org/content/10.1101/2024.04.10.588953v1....
I have read it now. Comment below. As to an increase in mouse neocortical volume—they have ZERO data. Look at their figures 1g and 1h and weep for the crap reviewing.
> can say with reasonsble assurance that no one has measured and weighed more mouse brains than I have
Out of curiosity how many would that be? Thousands? Tens of thousands? How long does it take to measure a mouse brain? When you do it does someone remove them for you or you also have to remove them before measuring?
Personal dissections about 2-3K (lifetime). Total over 35 years of the lab: about 20,000.
I guess that whatever recognises the enhancers must be conserved ish in mice for us to be able to drop this region into the mouse genome. That might be interesting alone apart from the functions of the frizzleds. Always a curious question about how much new “machinery” we have in humans, vs us being a specific configuration of common machinery (i.e., if we were to swap out introns/regulatory regions from a mouse with human, and assuming we don’t screw up any checkpoints, how far would it get with looking/acting like a human?).
Just reviewed the paper. Extremely elegant flamingo dance. But let me reserve my comments to the Hacker News theme of an enlarged brain (or really an enlarged neocortex). The key panels on this topic are Fig. 1g and 1h. OMG, the dissections of these "best examples" are sad–– the olfactory bulbs are serious torn and paraflocculi are detached. The histogram that is Fig 1h has a grand total of 6 measurement for the human transgenic mouse line and 4 (four) points for the control. Hey when you have a statistically significant result why mess it up by evaluating 10 of each of the two groups. You might not get published in Nature or airplay on HN, and would that not be a shame?
As for the measurement of brain size, it is a planimetric projection of the area of the dorsal surface of the neocortex. This is a lame way to do morphometry. Squish the brain a bit and the area will expand beautifully—and by way more than 6.5%.
One last comment on the genetics of these animals. The are probably incorrectly stated to be at least F8 progeny of the mixed 129-B6N embryonic stem cells. I hope they mean N8 backcross progeny––that is to say, 8th generation congenic lines. But in this case they appear to have backcrossed weirdly to a different type of B6; the standard C57BL/6J strain. All of this means that even in the best case, they have three different genomes banging around in supposedly co-isogenic cases and their controls: 1. chunks of 129 strain chromosomes that will still be common even at 8 generations, 2. chunks of B6N chromosomes that will also be common at 8 generations, and of course the B6J background strain. You would have to carry out sparse whole genome sequencing or use the GigaMUGA array to unconfound the genetics in this study.
> The key panels on this topic are Fig. 1g and 1h.
Is it possible that they are 2a and 2b in the preprint?
Since you mention both measuring and weighing, I'm now curious. How well does brain weight relate to brain size?
Highly linear in vivo and close to 1. Brain has a lot of fat but not a big deal in mice. Only fine point is that the density of brain varies depending on fixation protocol but almost always by less than 5% if you get the osmolarity of the solution just right.
I can't read the full paper...
Do they have a control group? (Bonus for a blind control group.)
Is the 6.5% statisticaly significant?
How many mice?
For the claim of increased neocortical size—-the topic of this HN thread, they show 4 controls versus 6 transgenic mice. Not sure if they did this blinded. Would depend mainly on when they did the genotyping.
The question remains if the same genes command an increase of the braincase, and what potential issues (including limited growth) happen if it's not the case
Meta comment: 11/13 comments currently are pop-culture references (Flowers for Algernon, Pinky and the Brain, other movies) or generic phrases and only 2 comments relate to the actual article. I've been on HN for over a decade, and this is the worst it's ever been.
Reading your comment now, 6 hours after you posted it. Most of those comments you reference are now dead and hidden.
Everything is fine, give it some time.
Comments still turn up useful insights that you don't usually catch on other places, like the top one. I think the trend of low quality AI related crap getting to top of the front page is more annoying.
Here's a formula you can try :
Title : "I made a AI powered insert tool use-case"
Article: "Hey look what I did ! By hooking up insert LLM model to this insert random tool in a few hours of vibe coding I got it to do something that vaguely resembles something useful and made no effort to evaluate performance/compare to anything. Here's my toy demo example."
Should be front page material, going on the last few months.
Worst is the rule that you're not allowed to make this comment. I think it actively allows it to slip further down into reddit level
Downvote, flag, and move on. These people don't understand the difference between HN and Reddit and there's always going to be more influx than you have bandwidth to talk out of it.
There are much better professional discussions on reddit, then in most HN threads. No need to be elitist.
> There are much better professional discussions on
That is the random peak that results from lucky combinations in some corners of a space - it is not owing to how the channel works, it is not owing to the culture it promotes.
Comparing the best of the best of Reddit against "most HN threads" isn't a good argument. Either compare the best of the best of both (I'd argue that HN comes out on top on most tech / programming related issues, and Reddit comes out on top more on more general topics), or compare averages to averages (Here HN comes out on top in my view).
The comparison of best-of-the best is not that useful in practice because you have to filter through so much noise to find it, so the average case is more useful in practice.
And after this comment, 2 out of 14 relate.
Agreed. But what can you do? People will comment what they will, and vote as they like.
Still could be worse, in a few years it'll just be ten comments all saying "Yeah, this is big brain time"
and all the smart people will be elsewhere. your job is to find elsewhere
Check out some of the timestamps in the links in last line of the guidelines. https://news.ycombinator.com/newsguidelines.html
I am aware of the guidelines, considering how long I've been here. However, I have offered proof, that over 80% of the comments do not contribute anything and are exactly like those of Reddit.
The enshittification of reddit and its consequences has been a disaster for discussion boards across the internet
Yes, for instance it sent me here.
Given the standard practice in lab experimentation using mice (which is to "dispose" of them after their usefulness for a test is done) I'm extremely troubled by the idea of gene editing experimentation using mice to make them more intelligent (their standard intelligence would already be surprising to many people).
While I already think it's unethical to use animals the way we do, the idea that we're experimenting with how close we can get to bestowing them with human intelligence before killing them is saddening.
That is neat but doesn't mean much by itself without some kind of test on cognition. The size of a creatures brain doesn't really correlate with intelligence very well.
Heck, if you give mice all of our DNA, they’ll become exactly like us!
But what if they get to keep their mouse DNA too? Do we get 2 meter large super smart mice? Sounds like a horror movie
So when you give the mouse some human DNA, do you then need to send the proposal to the Institutional Review Board, since these mice are a tiny bit human?
But are they any smarter?
They’re into crypto now.
> But are they any smarter?
So that's a "No" then... ;)[dead]
[flagged]
"The same thing we do every night..."
[flagged]
[flagged]
I like this better than sharks though, sharks are terrifying enough without super intelligence.
feel like if they were smarter they'd be far less scary. is not the main fear with sharks that they'll mistake you for something they actually want to eat?
[flagged]
[flagged]
The book is better, but... yeah, this was my first thought too.
[flagged]
[flagged]
comment on the comments geneticly spliced animals and humand are now real, but for many years have occupied a specific place in sci fi, where there are enhanced animals and humans, hybrids, and we have a huge subculture of "furryies" so the headline is pushing lots of buttons. The thing is that it's going to happen, and the pieces are here right now to make hybrids, and even alter a life forms DNA after birth. The point bieng if anybody is getting triggered by the whole kids and smart phones thing, best just think of that as quaint , as there is a massive market for some little adorable critter that can talk.....brace yourselves and nothing can make that possibility go away now
And thus began the Mousepocalypse...
(:shrug: I get it... Sore point... The apocalypse isn't a humorous thing to joke about, what with it being in the midst of actually happening and all ... but if I don't try to find myself a little giggle here and there once in a while I'ma literally go insane, so whatever. Down-vote away. Don't really care anymore.)