Yuchang Wu

@AlexTISYoung Claude is really good at software engineering. Tried both paid version of Gemini (edu version) and ChatGPT. I think Gemini is better for scientific research. But it needs very precise prompting, so I usually first ask it to optimize my prompt, then answer its own prompt.

Surprised Gemini is not getting much love given recent benchmark results. I wonder why?

Which AI pro subscription should I get for advanced technical/scientific research?

@SashaGusevPosts edu email account can get an edu version of Gemini, which should have less hallucinations.

There are still some pitfalls. It hallucinates in ways that I could see being a major time sink. I recently asked it to simulate stabilizing selection but it coded up directional selection in a way that would not have been obvious without already knowing the right math.

I've gradually started using Gemini as a kind of online notepad. By taking notes directly in the prompt it is then trivial to ask the AI to identify any immediate problems, do a lit review, implement a toy simulation, outline a talk, suggest ways to visualize the results, etc.

@Jiacheng_Miao very cool!

🚀 Introducing Paper2Agent - a system that turns static research papers into interactive AI agents you can chat with and use the tools/data in the paper. Just ask, “Run this paper’s method on my data.” The paper agent will handle the task for you. 1/n

@dougthespeed We first copy each chromosome in plink format to active environment, extract HM3 SNPs, and then merge all the HM3 SNPs from each chromosome. The copying process is slow (~1h for one larger chr?), but doable. You can run the job in background. Hail is always painful to me.

StatGen Question - Help with AllofUs data. We would like to extract genotypes for all individuals and ~1M HapMap3 SNPs in PLINK (1) format (which we can then use for future analyses). Has anyone done this and can offer advice?

@AlexTISYoung sorry to hear this. Hope you'll have a speedy and full recovery!

In March last year, at the age of 35, I was diagnosed with advanced stage III rectal cancer with a metastasis in my liver. This was a shock: I had no family history, and none of the doctors suspected it. In fact, I'd had a negative occult blood test when I went to ER with severe digestive symptoms a few months before. The first lesson I learnt from this was that there's no substitute for the proper diagnostic procedures: if you have any bleeding, get a colonoscopy, even if you don't have family history. I was treated with total neoadjuvant therapy at UCLA. I had 6 weeks of combination radiation and chemotherapy (Xeloda). The photo shows me after completing that first step of my treatment along with my parents in Yosemite. My mother has already survived two breast cancers, and my father had recently had a knee replacement (at the age of 70), so getting all three of us to the top of Sentinel Dome felt like an achievement. The chemoradiation was followed by 6 rounds of combination chemotherapy: oxaliplatin infusions followed by two weeks of Xeloda. I tolerated this unusually well. During this period, I managed to hike up to over 11,000ft multiple times in the Sierras, and I wrote the main text of a 72 author meta-analysis paper and submitted it to Nature where it has passed first round reviews (currently working on revisions). I showed a good clinical response, with the tumor in my rectum nearly completely disappearing and the liver metastasis shrinking substantially. At the end of October I had surgery: I was under for 9 hours, and they cut out a large bit of my rectum, 1/3rd of my liver, and gave me a temporary ileostomy. Waking up from this was probably the strangest experience of my life: I felt more machine than man, and I had wild hallucinations form such a large dose of anesthesia. It took a while to feel OK after that surgery. However, I still had the ostomy. I had that reversed at the end of January in another surgery, which took a greater toll on me than I expected. There's a cumulative effect of having so many major medical interventions, and it reveals anything weak in body or mind. Surprisingly, my mind held up well during this process. This was thanks to the love and support I received from friends and family and the great care I received from UCLA — in particular Dr Anand (medical oncology), Dr Kazanjian (colorectal surgery), Dr Agopian (liver surgery), and the radiation oncology team. I only just started to feel like I was starting to feel OK again and I got bad news: I got a positive circulating tumor DNA test (ctDNA), the signatera test from natera. Unfortunately, the second test I had done recently showed the level of ctDNA in my blood is increasing. This means disease recurrence is almost certain, likely within a year from the first positive test. Since then I've been put on celecoxib based on recent data indicating this can reduce risk of disease recurrence, although it doesn't seem to agree to well with my digestion. I've had CT, MRI, colonoscopy and there's been nothing to see, but it's only a matter of time. My oncologist thinks it is very likely (90%+) that it will be a local recurrence in my liver, which should be fairly easy to cure. But there's a small chance it is something worse, even potentially incurable. This is the reality of cancer for many patients: years of uncertainty. It's still not clear why this happened. Genetic testing returned nothing. My polygenic risk score (PRS) — something close to my own research — gave me totally average risk, at least according to 23andMe. I've always been slim, fit, eaten a pretty healthy diet. I was even raised vegetarian by hippy parents. I probably drank and partied more than is medically advised, but nothing extreme. However, there's been a well-documented uptick in cases like mine. A recent paper indicated this may be due to colobactin, a bacterial mutagen associated with E. coli among other bacteria. I may look at my Tempus tumor data to see if the somatic mutation in the APC gene they found (the only driver mutation) has a signature matching colobactin. If anyone knows anyone with a worthwhile expert opinion on how to manage my situation I'd be interested to hear!

@SashaGusevPosts good question. I don't know since I'm not closely following on this topic.

@yy_stat Thanks, I hadn't seen this before, great commentary. Has anyone done the analysis of real couples since then?

The New York Times recently had an article on IVF and embryo selection which I think buries a few important ledes about these products ...

update: The filter function on data browser does not give accurate results: some categories and some variants meet the criteria are missing. Guess we have to use VAT for more serious work.

Wow! Need a LoF variant list in AllofUs. Filter the giant variant annotation table? Hours of waiting! Ended up copying and pasting data browser results to DeepSeek and in 10 seconds I got a clean tab-separated file! Cool!

@SashaGusevPosts Does the environment correlate with genetics? Otherwise, I don't have the intuition why GWAS could be problematic.

Run a GWAS that doesn't control for population stratification. The GWAS will orient all of the drifted alleles towards the environmental differences and the resulting polygenic score will massively exaggerate population genetic differences that do not actually exist.

How population stratification makes environments look like genes. A short 🧵:

My first co-author paper is now preprinted! With @ZijieZhao1996 and @Q_StatGen we introduce PUMAS-ensemble, a regularized ensemble PRS requiring only GWAS summary statistics and LD reference data as input. biorxiv.org/content/10.110…

Our new PRS paper is out! We introduce a method to create regularized ensemble PRS on GWAS sumstats. Individual-level data is no longer needed for sophisticated ensemble learning @ZijieZhao1996 @stphn_drn Paper📰:biorxiv.org/content/10.110… Software🧑‍💻:github.com/qlu-lab/PUMAS

One score to rule them all: regularized ensemble polygenic risk prediction with GWAS summary statistics biorxiv.org/cgi/content/sh… #biorxiv_genetic

@geyu_zhou yes; I was too excited to see the job done and totally forgot to copy the results to bucket

@yy_stat Using dsub and writing to the workspace bucket is a safer approach.

pain in using cloud computing: took 1 day 10 hours and 44 minutes to finish some analysis using AllofUs data, and I accidentally deleted the "environment" and all the results are gone forever ... 😭😭😭

@dajiangliu81 @AllofUsResearch @AllofUsCEO email them and they're usually very responsive. hope this could you resolved without any problems!

@AllofUsResearch @AllofUsCEO Want to bring this to your attention. We set up a workspace without linking a GCP account. We are doing an extensive computational project. No one notified us anything until yesterday and we were told that our data would be deleted in 24 hours. 1/n

#ASHG24 Interested in how to reliably use AI/ML to accelerate genetic discovery? Stop by poster 4151 this afternoon to chat!

Four presentations from us this year, covering very diverse topics. Stop by and say hi🍷#ASHG24 #ASHG2024 1) Genetic effect heterogeneity @yy_stat 2) Genetics of partner choice @S5b6bY284u9qqoa 3) Perturb-seq/GWAS data integration @stphn_drn 4) ML-assisted GWAS @Jiacheng_Miao

While we didn't explore it in this paper, another tool (POP-GWAS) from our lab (x.com/Q_StatGen/stat…) may offer a better solution when dealing with phenotypes with biases. --- also check it out!

Thrilled to see this published on Nature Genetics! We found weird results using AD GWAX in another project and discovered that the bias is actually a very serious issue in the AD field. This underscores the importance of data quality. Larger is not always better!