Nikita Vaulin

Excited to share #AlphaGenome, a start of our AlphaGenome named journey to decipher the regulatory genome! The model matches or exceeds top-performing external models on 24 out of 26 variant evaluations, across a wide range of biological modalities.1/6

@maria_spies @PietroPichierri Congrats!🎉🎉

It's out! nature.com/articles/s4158… Structure of the two RAD52 rings forming a spool-like structure at replication fork, rearranging and protecting the fork. Tour-de-force work by Masa Honda and Ali Razzaghi and wonderful (as usual) collaboration with @PietroPichierri 's lab.

Happy Polycomb Day (3/27)! Today we celebrate H3K27me3- Polycomb’s quiet mark of repression, where gene silencing is not neglect, but a deliberate act of identity.

Computational PhD Opportunity! Funded by @FWF_at Come and work with us to discover the early pediatric tumor initiation events with innovative iPSC models and single-cell omics! At @StAnna_CCRI! ccri.at/working-at-the…

Today, he made a force push for massive changes to the lab's website. Don't fall, don't fall, don't fall...

@LGentzbittel Thank you, Prof. Gentzbittel! I'm really happy that the A ∩ B > 0 (where A is your 4 years in Sk and B is mine 2 years in Sk).

Today marks my last day at Skoltech, and what a journey it has been during these 4 years ! Grateful for the incredible experience and proud of what we’ve achieved together:

@leocastorina @ktv_xn, your turn

#Protein People 🧬✨: What is your favourite protein and what makes it so special to you? Bonus points for designed proteins! I’ve got an exciting project brewing that I can’t wait to reveal soon! Drop your answers below and stay tuned 👀🎉

@shitov_happens @stethospeaks I have the same issue, but with SCENIC and french auto industry

@stethospeaks I discovered this earlier because every time I googled the package, the art would appear

Can't believe that I had to visit the Art Institute of Chicago to see Georges Seurat's paintings in distinctive pointillism to realize that the single-cell RNA-seq analysis package in R is named Seurat because the appearance of cell clusters resembles the artist's style. 🫠

@Sanbomics As you say union looks better, but has higher std because some particular methods are not good on some particular data - maybe it worth trying something in between to integrate predictions. For instance, a majority vote.

Are two single-cell doublet annotation methods better than one? Sometimes, but unfortunately, five are even better... I test five different methods, their union, and their intersection in 23 hashtag datasets. Average metrics shown below: (1/6)

(1) How can we build complex structures from a limited set of progenitors and cell types? In our new preprint, we describe patterns of cell fate decisions during early development and (in some cases) identify molecular correlates of different behaviors. biorxiv.org/content/10.110…

@sokrypton @chrisfrank662 @DominikSchiwie0 @hendrik_dietz Hilarious! The noble gift for a colleague is a necklace with a small glass flask and inside is a set of proteins with his/her name

A few more letters... and we can finally write amino acid sequence with protein structure encoded by amino acid sequence? 🧬🤔😀 @chrisfrank662 @DominikSchiwie0 Lara Fuss @hendrik_dietz biorxiv.org/content/10.110…

I debuted on Medium with my heartfelt cry regarding the aging clocks. The article briefly outlines the main result of our recent article, interspersed with my personal concerns and feelings for the field of aging biology. Enjoy reading! @firepaladiner/whats-wrong-with-aging-clocks-2b82db40d853" target="_blank" rel="nofollow noopener">medium.com/@firepaladiner…

@flanditto @hendersi Yeah, it should be some regulatory play. However, Met is not the only one start codon. There are some "non-canonical" starts (which is interesting itself btw)

Another question from a student that I don't have a good answer for (chatgpt also struggled) - in the genetic code why do some amino acids have only one codon (eg Met, Trp), whereas others get six (eg Ser)?

@lampimampi Добавлю в копилку список незаменимых аминокислот: ЛИЗа МЕТнула ФЕН в ТРИбуну, ТРЕзвый ЛЕЙтенант ВАЛялся в ИЗОляторе с АРГентинским ГИтаристом Правило прекрасно тем что оно идет не по первой букве, а по первым трем - ровно так как все и обозначают аминокислоты

Какое лучшее мнемоническое правило вы знаете? Моё любимое — список основателей первой в России марксистской организации «Освобождение труда», который советские студенты запоминали по звучному акрониму: — Плеханов — Игнатов — Засулич — Дейч — Аксельрод

@J_my_sci Oh, sorry, the one that was below in the thread. Yeah, that's hard to do, i see

@J_my_sci Nice work! In the paper you show the homogeneity of i3Neurons, but have you tried to compare proteomes of your iPSC-derived neurons and human neurons?

@RobKlose @AleksSzczurek @emmy_dimitrova @Jess_R_Kelley @blackledge_neil Wow, that's great work! Now I wonder if somehow these enzymatic dynamics change during cell differentiation with different PRC1 assemblies.

Great to see @AleksSzczurek's paper out today! A wonderful team effort with @emmy_dimitrova , @Jess_R_Kelley , and @blackledge_neil. Have a look if you are interested in how #Polycomb controls transcription to repress gene expression. rdcu.be/dTD3t

@FabPrewett @JoachimSchork Here is an example for bell-shaped distributions, which I am sure can confuse anyone. But with the QQ-plot the deviation from normality is clearly visible. Moreover, you can make a QQ-plot for any distribution, not only normal.

@JoachimSchork Could you not just plot the distribution to see if it is normal or not?

Evaluating the normality of your data is crucial in statistical analysis, as many techniques assume that the data follows a normal distribution. The QQ plot (Quantile-Quantile plot) is a simple yet powerful tool to visually assess this. ✔️ Improved Model Accuracy: Properly handled, QQ plots help confirm that your data meets the assumptions required for many statistical methods, leading to more reliable results. ✔️ Enhanced Data Insights: By detecting deviations from normality, you can make informed decisions about data transformations or the choice of statistical methods. ❌ Misleading Results: If normality is not checked, you may apply inappropriate statistical techniques, leading to incorrect conclusions. ❌ Overlooked Data Issues: Failing to evaluate normality might cause you to miss out on important patterns or outliers in your data. When interpreting a QQ plot, you are looking for how closely the points follow the reference line. 🔹 Normal Distribution: If the data follows a normal distribution, the points will align closely with the diagonal reference line. 🔹 Non-Normal Distribution: If the points significantly deviate from the line, this indicates that the data is not normally distributed. The nature of the deviation can provide insights into whether the data is skewed, has heavy tails, or other issues. The visualization in the post contrasts two QQ plots: the left plot shows a data set following a normal distribution, where the points align closely with the reference line. The right plot displays a data set with non-normal distribution, where points deviate significantly from the line. Curious about how to create QQ plots in R? Check out this tutorial: statisticsglobe.com/r-qqplot-qqnor… If you want to learn more about statistical methods in R, check out my online course on Statistical Methods in R, starting on September 9, 2024, which covers this topic and others in further detail. Learn more: statisticsglobe.com/online-course-… #datavisualization #rprogramminglanguage #programming #datastructure #visualanalytics #database

@pedrotomazsilva This is interesting! Looks similar to Hi-C maps. And your fig. 6A plot is called a "scaling plot" in 3D genomics, for example. Have you tried to compare/align your matricies with the Hi-C matricies of the same DNA regions from any public data?

Have you ever wondered what the genome looks like through the eyes of a DNA language model? In our newest preprint we use DNA LMs to study nucleotide dependencies in the genome, revealing functional elements, characterizing variants and evaluating DNA LMs tinyurl.com/6wbwjaf4

@byu_sam Wow, insane! Interestingly, the first author in this sc-proteomics paper is an otolaryngology doctor (sechenov.ru/univers/about_…). And all Sechenov-affiliated authors can be found in russian dissernet database (which collects works with purchased co-authorship, plagiarism etc)

Well. It REALLY happened. The plagiarized paper got published. my paper from 2021 - pubmed.ncbi.nlm.nih.gov/33915259/ new one from 2024 - pubmed.ncbi.nlm.nih.gov/38700052/ We'll look at figures in the 🧵 below.