Alex Rubinsteyn
@alexr.bsky.social
personalized cancer immunotherapy = genomics + immunology + machine learning + oncology
(pirl.unc.edu)
(pirl.unc.edu)
Got stuck on a few different experiment design queries with ChatGPT 5 Pro
Finally got the Gemini app and tried Gemini 3 and…it perfectly answered my question with every relevant experimental detail across a few different papers/patents
Finally got the Gemini app and tried Gemini 3 and…it perfectly answered my question with every relevant experimental detail across a few different papers/patents
November 25, 2025 at 12:54 AM
Got stuck on a few different experiment design queries with ChatGPT 5 Pro
Finally got the Gemini app and tried Gemini 3 and…it perfectly answered my question with every relevant experimental detail across a few different papers/patents
Finally got the Gemini app and tried Gemini 3 and…it perfectly answered my question with every relevant experimental detail across a few different papers/patents
Re #2 -- how much did care improve around administration of ICIs over the past 10 years? Well, if you just look at the survival curves from this paper stratified by start year, you see monotonic increase of survival asymptotes from ~2016-2020
October 30, 2025 at 3:46 PM
Re #2 -- how much did care improve around administration of ICIs over the past 10 years? Well, if you just look at the survival curves from this paper stratified by start year, you see monotonic increase of survival asymptotes from ~2016-2020
I think there is still some residual evidence of the claimed interferon effect from the paper but it's small & cohort size isn't big enough to really tease it out confidently.
Only with the wrong comparison group can you go from p=0.270 to "published in Nature"
Only with the wrong comparison group can you go from p=0.270 to "published in Nature"
October 30, 2025 at 3:37 PM
I think there is still some residual evidence of the claimed interferon effect from the paper but it's small & cohort size isn't big enough to really tease it out confidently.
Only with the wrong comparison group can you go from p=0.270 to "published in Nature"
Only with the wrong comparison group can you go from p=0.270 to "published in Nature"
Anyways, re: #1 / #2 -- the effect of just splitting out early ICI patients from later ones gives you pretty much all the claimed effect (without the gymnastics around using vaccination status to pull out the later patients)
October 30, 2025 at 3:35 PM
Anyways, re: #1 / #2 -- the effect of just splitting out early ICI patients from later ones gives you pretty much all the claimed effect (without the gymnastics around using vaccination status to pull out the later patients)
Tried going back to Claude today and it’s…totally useless for biomed? Got this for many different questions in a row but can eventually get an answer if I write a long shpiel about why it’s really safe to tell me about an IHC CRO.
September 26, 2025 at 2:20 PM
Tried going back to Claude today and it’s…totally useless for biomed? Got this for many different questions in a row but can eventually get an answer if I write a long shpiel about why it’s really safe to tell me about an IHC CRO.
Very preliminary figure from Steve Vensko in our lab but still wanted to share it.
% melanoma patient samples (n=27) expressing CTA
vs.
% cells in non-zero samples expressing the CTA
% melanoma patient samples (n=27) expressing CTA
vs.
% cells in non-zero samples expressing the CTA
September 20, 2025 at 2:18 AM
Very preliminary figure from Steve Vensko in our lab but still wanted to share it.
% melanoma patient samples (n=27) expressing CTA
vs.
% cells in non-zero samples expressing the CTA
% melanoma patient samples (n=27) expressing CTA
vs.
% cells in non-zero samples expressing the CTA
Is there something wrong with the HLA-II entries of the HLA Ligand Atlas?
Looked up the model PRAME epitope A2/SLLQHLIGL and found it's mostly detected on HLA-II alleles -- which doesn't make sense of a teeny little 9mer.
(it's also weird to see it so much on healthy samples)
Looked up the model PRAME epitope A2/SLLQHLIGL and found it's mostly detected on HLA-II alleles -- which doesn't make sense of a teeny little 9mer.
(it's also weird to see it so much on healthy samples)
September 18, 2025 at 10:54 AM
Is there something wrong with the HLA-II entries of the HLA Ligand Atlas?
Looked up the model PRAME epitope A2/SLLQHLIGL and found it's mostly detected on HLA-II alleles -- which doesn't make sense of a teeny little 9mer.
(it's also weird to see it so much on healthy samples)
Looked up the model PRAME epitope A2/SLLQHLIGL and found it's mostly detected on HLA-II alleles -- which doesn't make sense of a teeny little 9mer.
(it's also weird to see it so much on healthy samples)
Hominid re-diversification
September 18, 2025 at 10:34 AM
Hominid re-diversification
Elementary schools teach non-commutative integer multiplication!?
September 16, 2025 at 2:51 PM
Elementary schools teach non-commutative integer multiplication!?
Comfy-ness monotonically decreasing
July 24, 2025 at 7:56 PM
Comfy-ness monotonically decreasing
Why is mass spec tooling all like this?
June 4, 2025 at 4:00 PM
Why is mass spec tooling all like this?
Wow, ~$1B cut from program to identify and track pathogen outbreaks. Seems like this is their whole budget?
@helenbranswell.bsky.social
-- what happens now if there's a novel pathogen outbreak, do local health labs not have any funding to characterize and track it?
@helenbranswell.bsky.social
-- what happens now if there's a novel pathogen outbreak, do local health labs not have any funding to characterize and track it?
June 4, 2025 at 3:42 PM
Wow, ~$1B cut from program to identify and track pathogen outbreaks. Seems like this is their whole budget?
@helenbranswell.bsky.social
-- what happens now if there's a novel pathogen outbreak, do local health labs not have any funding to characterize and track it?
@helenbranswell.bsky.social
-- what happens now if there's a novel pathogen outbreak, do local health labs not have any funding to characterize and track it?
June 4, 2025 at 3:28 PM
It's bad to take the smartest and most curious young people in our society and funnel them into a decade of low pay apprenticeship where they slowly learn the arcane levers of prestige ritual and funding bureaucracy so they can have a better chance of becoming grant gremlins.
June 2, 2025 at 3:29 PM
It's bad to take the smartest and most curious young people in our society and funnel them into a decade of low pay apprenticeship where they slowly learn the arcane levers of prestige ritual and funding bureaucracy so they can have a better chance of becoming grant gremlins.
I know you all have already collectively shrugged it off but holy cow: we really solved protein folding
(shown: AF3 with 3 chains of a PRAME bispecific all cuddled up in the right spots, you can see His tag tail on the bottom & (GS^4)^4 linker between two parts of the backbone)
(shown: AF3 with 3 chains of a PRAME bispecific all cuddled up in the right spots, you can see His tag tail on the bottom & (GS^4)^4 linker between two parts of the backbone)
May 29, 2025 at 7:40 PM
I know you all have already collectively shrugged it off but holy cow: we really solved protein folding
(shown: AF3 with 3 chains of a PRAME bispecific all cuddled up in the right spots, you can see His tag tail on the bottom & (GS^4)^4 linker between two parts of the backbone)
(shown: AF3 with 3 chains of a PRAME bispecific all cuddled up in the right spots, you can see His tag tail on the bottom & (GS^4)^4 linker between two parts of the backbone)
If you generate an AI image and then feed as the input to generate another one, actually the process always converges in this fixed point. No one knows why.
May 28, 2025 at 10:50 AM
If you generate an AI image and then feed as the input to generate another one, actually the process always converges in this fixed point. No one knows why.
We're still waiting for another round of experiments to fully resolve this but here are some preliminary results:
{mRNA, peptide + CpG} >>> {peptide + polyIC, peptide + montanide}
STING agonist does not significantly enhance CpG
Upcoming data with poly-ICLC +/- montanide
{mRNA, peptide + CpG} >>> {peptide + polyIC, peptide + montanide}
STING agonist does not significantly enhance CpG
Upcoming data with poly-ICLC +/- montanide
May 16, 2025 at 2:50 PM
We're still waiting for another round of experiments to fully resolve this but here are some preliminary results:
{mRNA, peptide + CpG} >>> {peptide + polyIC, peptide + montanide}
STING agonist does not significantly enhance CpG
Upcoming data with poly-ICLC +/- montanide
{mRNA, peptide + CpG} >>> {peptide + polyIC, peptide + montanide}
STING agonist does not significantly enhance CpG
Upcoming data with poly-ICLC +/- montanide
North Carolina NUT carcinoma symposium was really successful. Lots of scientific mingling, felt like a bigger research community coming together.
Last session was a future therapeutics brainstorm -- hope the Vincze family's nascent NUT Carcinoma Alliance can fund some of these as trials.
Last session was a future therapeutics brainstorm -- hope the Vincze family's nascent NUT Carcinoma Alliance can fund some of these as trials.
March 9, 2025 at 12:22 AM
North Carolina NUT carcinoma symposium was really successful. Lots of scientific mingling, felt like a bigger research community coming together.
Last session was a future therapeutics brainstorm -- hope the Vincze family's nascent NUT Carcinoma Alliance can fund some of these as trials.
Last session was a future therapeutics brainstorm -- hope the Vincze family's nascent NUT Carcinoma Alliance can fund some of these as trials.
Wish we had a more faithful model of mRNA-LNP reactogenicity in mice -- the IL-1 vs. IL-1ra mismatch means they don't lose weight and mRNA gets to live in its own magic quadrant on this plot of T-cell response vs. weight change from vaccination
February 28, 2025 at 8:25 PM
Wish we had a more faithful model of mRNA-LNP reactogenicity in mice -- the IL-1 vs. IL-1ra mismatch means they don't lose weight and mRNA gets to live in its own magic quadrant on this plot of T-cell response vs. weight change from vaccination
CoT vs answer
(Has anyone de-aligned R1 yet?)
(Has anyone de-aligned R1 yet?)
January 25, 2025 at 2:28 AM
CoT vs answer
(Has anyone de-aligned R1 yet?)
(Has anyone de-aligned R1 yet?)
New de-novo TCR design preprint from @dkarthikey1.bsky.social just dropped:
TCR-TRANSLATE: Conditional Generation of Real Antigen Specific T-cell Receptor Sequences
www.biorxiv.org/content/10.1...
TCR-TRANSLATE: Conditional Generation of Real Antigen Specific T-cell Receptor Sequences
www.biorxiv.org/content/10.1...
November 13, 2024 at 3:29 AM
New de-novo TCR design preprint from @dkarthikey1.bsky.social just dropped:
TCR-TRANSLATE: Conditional Generation of Real Antigen Specific T-cell Receptor Sequences
www.biorxiv.org/content/10.1...
TCR-TRANSLATE: Conditional Generation of Real Antigen Specific T-cell Receptor Sequences
www.biorxiv.org/content/10.1...
I don't think I'm winning any graphic design prizes.
November 4, 2024 at 8:06 PM
I don't think I'm winning any graphic design prizes.
You probably already know this but use CEFX (not CEF) as your positive control for Tcell assays
November 1, 2024 at 1:26 AM
You probably already know this but use CEFX (not CEF) as your positive control for Tcell assays
