Coleen Murphy
@ctmurphy1.bsky.social
Scientist, citizen, mom, and occasionally athletic
Summary: THIS AI REVIEWING TOOL IS DANGEROUS. If editors use AI tools to review, they may desk-reject papers just because the tool suggests many but pointless experiments, while missing fundamental flaws.
January 15, 2026 at 4:51 PM
Summary: THIS AI REVIEWING TOOL IS DANGEROUS. If editors use AI tools to review, they may desk-reject papers just because the tool suggests many but pointless experiments, while missing fundamental flaws.
b. Our paper: the review tool missed the whole point of the paper (memory) and suggested about 3yrs worth of pointless experiments (all were on lifespan/healthspan, not at all the point of the paper).
January 15, 2026 at 4:51 PM
b. Our paper: the review tool missed the whole point of the paper (memory) and suggested about 3yrs worth of pointless experiments (all were on lifespan/healthspan, not at all the point of the paper).
a. JC-reviewed paper: our lab's JC review identified a fatal, fundamental flaw in the paper that rendered the whole study useles, as well as other more minor problems layered on top of that, but the review tool only identified the minor issues, and missed the major problem with the whole paper.
January 15, 2026 at 4:51 PM
a. JC-reviewed paper: our lab's JC review identified a fatal, fundamental flaw in the paper that rendered the whole study useles, as well as other more minor problems layered on top of that, but the review tool only identified the minor issues, and missed the major problem with the whole paper.
4. AI reviewing tools: This was most eye-opening to me. Submitted two bioRxiv papers we knew well - one we had a journal club on to review, the other from our lab - to the new open-access reviewing tool available through bioRxiv. Results were terrible in both cases:
January 15, 2026 at 4:51 PM
4. AI reviewing tools: This was most eye-opening to me. Submitted two bioRxiv papers we knew well - one we had a journal club on to review, the other from our lab - to the new open-access reviewing tool available through bioRxiv. Results were terrible in both cases:
3. AI and fraud: terrifyingly good facsimiles of Western blot data. Not sure how journals are going to deal with this.
January 15, 2026 at 4:51 PM
3. AI and fraud: terrifyingly good facsimiles of Western blot data. Not sure how journals are going to deal with this.
2. What AI is not OK/very bad for:
a. Writing - we are trying to teach students to write, so if they use AI, they are not learning the material and they are not learning how to write well.
a. Writing - we are trying to teach students to write, so if they use AI, they are not learning the material and they are not learning how to write well.
January 15, 2026 at 4:51 PM
2. What AI is not OK/very bad for:
a. Writing - we are trying to teach students to write, so if they use AI, they are not learning the material and they are not learning how to write well.
a. Writing - we are trying to teach students to write, so if they use AI, they are not learning the material and they are not learning how to write well.
Good to see people out protesting against ICE - and lots of support from drivers passing (the exceptions were notable and few)
January 11, 2026 at 7:11 PM
Good to see people out protesting against ICE - and lots of support from drivers passing (the exceptions were notable and few)
This is a ridiculous place to sleep
December 6, 2025 at 7:05 PM
This is a ridiculous place to sleep
The Princeton museum opening was (is) a blast! Best costumes: Frida Kahlo (several), Louvre heist crew
November 1, 2025 at 4:36 AM
The Princeton museum opening was (is) a blast! Best costumes: Frida Kahlo (several), Louvre heist crew
Saw a nice moon through a telescope tonight
October 28, 2025 at 5:47 AM
Saw a nice moon through a telescope tonight
Andrew Zwicker - physicist & Representative
October 18, 2025 at 7:21 PM
Andrew Zwicker - physicist & Representative
Checking in from the Princeton NJ No Kings rally
October 18, 2025 at 7:12 PM
Checking in from the Princeton NJ No Kings rally
My littlest turns 15 today!
September 21, 2025 at 4:15 PM
My littlest turns 15 today!
Following directions is hard
June 18, 2025 at 3:36 AM
Following directions is hard
No Kings protest in Princeton NJ
June 14, 2025 at 7:35 PM
No Kings protest in Princeton NJ
Apparently this is now Athena's favorite place to sleep and Dash's favorite chew toy - meanwhile the scratching post hasn't been touched
Also: Abolish ICE
Also: Abolish ICE
June 8, 2025 at 3:11 PM
Apparently this is now Athena's favorite place to sleep and Dash's favorite chew toy - meanwhile the scratching post hasn't been touched
Also: Abolish ICE
Also: Abolish ICE
Of course the box is better than the toy
June 7, 2025 at 9:17 PM
Of course the box is better than the toy
Jane Boyer read and annotated my book cover to cover! So impressive
June 5, 2025 at 8:10 PM
Jane Boyer read and annotated my book cover to cover! So impressive
Our model is that non-autonomous regulation of memory might be a good way to extend the memory function in times of stress on a longer timescale than neurons might provide, paralleling the effects seen from blood factors that are secreted from the liver that improve cognition (Horowitz, Villeda).
June 4, 2025 at 8:58 PM
Our model is that non-autonomous regulation of memory might be a good way to extend the memory function in times of stress on a longer timescale than neurons might provide, paralleling the effects seen from blood factors that are secreted from the liver that improve cognition (Horowitz, Villeda).
So why the hypodermis? We think that the hypodermis is not just skin; it is also a highly metabolic tissue, with a transcriptome that best matches the human liver and blood plasma.
June 4, 2025 at 8:58 PM
So why the hypodermis? We think that the hypodermis is not just skin; it is also a highly metabolic tissue, with a transcriptome that best matches the human liver and blood plasma.
Finally, we see that aged daf-2 mutants rely on CREB almost entirely for their extended memory span, so there is a shift from neurons using CREB-independent to CREB-dependent targets to extend memory.
June 4, 2025 at 8:58 PM
Finally, we see that aged daf-2 mutants rely on CREB almost entirely for their extended memory span, so there is a shift from neurons using CREB-independent to CREB-dependent targets to extend memory.
We found that CREB is required for about half of daf-2's short-term memory extension, and for all of hyp DAF-2 and OSM-11 overexpression memory extension
June 4, 2025 at 8:58 PM
We found that CREB is required for about half of daf-2's short-term memory extension, and for all of hyp DAF-2 and OSM-11 overexpression memory extension
When we used snSeq to figure out the signaling, we found that the CREB transcription factor is broadly upregulated in neurons, including “memory” neurons, as are CREB’s targets.
June 4, 2025 at 8:58 PM
When we used snSeq to figure out the signaling, we found that the CREB transcription factor is broadly upregulated in neurons, including “memory” neurons, as are CREB’s targets.