Garth Kong | Computational Biologist

Computational Biologist @OHSUKnight in 🌲 Portland, OR 🌲 at Olga Nikolova's lab. Currently hacking into the m̶a̶i̶n̶f̶r̶a̶m̶e̶ genome 💻

[Garth Kong | Computational Biologist @KongGarth]

How reproducible are #bioinformatic pipelines really? 🤔

In pursuit of this question, I asked three lab members to run my pipeline on the same CUT&Tag data (4 conditions, 3 biological replicates, and 2 epigenetic marks). 💻

#Genomics #snakemake #AcademicChatter

[Mitch Guttman @mitchguttman]

Gene expression involves thousands of proteins that bind DNA, yet comprehensively mapping these is challenging. We developed ChIP-DIP – a method for simultaneous, genome-wide mapping of hundreds of DNA-protein interactions in a single experiment. https://t.co/NEH92mQTGs

[Ben Siranosian @BenSiranosian]

Been reading a lot of pathology AI papers lately... has got me feeling like:

[Garth Kong | Computational Biologist @KongGarth]

What do you call someone who pretends to know everything about epigenetics?

A chromatin charlatan

#epigenetics #bioinformatics

[Hussein Abbas @HusseinAbbasLab]

Hierarchy composition in AML matters. Our work on the epigenetic accessibility profiles of these hierarchies is now out. https://t.co/OG1w40Uxkr @MDAndersonNews @LeukemiaRF

[Garth Kong | Computational Biologist @KongGarth]

Speaker: because of the diff exp results, it supports my hypothesis 😁
Listeners: but what if you recluster it with different parameters? 🤔
Speaker: we don't talk about that here 🙁

[Patrick Hsu @pdhsu]

scientist after sequencing lots of single cells

[Ed Yong is not here @edyong209]

Some news: After 8 years & 750 stories, I've decided to leave The Atlantic. Today's my last day.
Being a writer means you can’t say things like "I can’t tell you what this means" cos, well, I can. That's kind of the point of me. So here’s an attempt at looking back & forward: 1/

[Scott Tyler (@[email protected]) @ScienceScottT]

Yes, tSNE and UMAP overfit the data creating the impression of structure from nothing (even with 1000 obs & 200 features, not 2 -> 2). I don't have a twitter 'team' that I'm fighting for/against, but I had a hypothesis, tested it, and I'm just reporting the results...

[Garth Kong | Computational Biologist @KongGarth]

I started to study #biology ten years ago. What nobody told me was that it (serendipitously) transforms into linear algebra, multivar statistics, and Greek 😁

#Bioinformatics #math

[Graham Heimberg @gheimberg]

Ever analyze a scRNAseq dataset and wonder if a specific cell state has been seen before? And if so, where in the human body? Under what conditions? Well, now you can use our lightning fast SCimilarity search and foundational model for that! ⚡️🔎🧬 https://t.co/t8L1xFOffd (1/11)

[Garth Kong | Computational Biologist @KongGarth]

This is actually very creative 😁

[Alicia Oshlack @AliciaOshlack]

Make sure you grab a snack BEFORE you check out ⁦@LonsBio⁩ posters #253 and #254 at #icg2023 🥑🧀🍇🍅🍫

[Claudio Cantù @ClaudioCantu81]

Have you wondered why the overlap between ChIP-seq or CUT&RUN replicates is often so low?

Should one only trust the small overlapping set?

@AnnaNordin96 @ppagella86 @GianlucaZamba and I tried to solve this conundrum central to signal detection theory 📈
https://t.co/Zz6woBHsVw

[Garth Kong | Computational Biologist @KongGarth]

@HarmitMalik Steve Henikoff = HHMI scientist during the day, street artist at night

[Garth Kong | Computational Biologist @KongGarth]

@jake_vancampen wow this is awesome!

[Johannes Köster (@johanneskoester.bsky.social) @johanneskoester]

#Snakemake 7.28 has been released. @jake_vancampen and @katatonikkat have contributed native support for @azure batch execution. Also many thanks to @microsoft for providing a free account for continuous testing.

[Garth Kong | Computational Biologist @KongGarth]

@Gurkan_Yardimci Took me a while to find but here it is:

https://t.co/VQKBKEqT2q

I noticed it has lots of topology sections which may not be directly to our work, so just gotta pick and choose

[Garth Kong | Computational Biologist @KongGarth]

Girl: You're so interesting, so what else are you into?
Me: Recently I'm getting into the #EDM scene nbd 😏
EDM:

[Chirag Nepal @ChiragNepal]

Do most #promoters have #H3K4me1 marks? YES !?
Do most #enhancers have #H3K4me3 marks? NO !?

Enhancers and promoters have different sequence composition #CGdensity.

Promoters are GC-rich and H3K4me3 binds to accessible CpG regions.

[Garth Kong | Computational Biologist @KongGarth]

@martinalexsmith That's over 1300 genomes per day... better get started quickly 😀

[Garth Kong | Computational Biologist @KongGarth]

@schonrockz Not pharm, but if you are analyzing processed/semi-processed data (e.g. tables) then R is preferred bc of its comprehensive stats library and data manip pkgs. Python is often useful for general processes like downloading + cleaning files. Does this help? Good luck!

[Garth Kong | Computational Biologist @KongGarth]

Thank YOU @johanneskoester for continuing to develop new features and fix bugs all these years!

[Johannes Köster (@johanneskoester.bsky.social) @johanneskoester]

Updated usage statistics for #Snakemake: we are now at >10 new citations per week on average and over 700k downloads of the #Conda package (not even counting pypi). Thanks a lot for the active use and all the feedback and contributions! https://t.co/WwE4GKpryY

[Charles Swanton @CharlesSwanton]

Today our important study on air pollutants in the promotion of lung cancer has been published in @Nature led by @WillHilliam @LimEmilia @DrClareWeeden @SwantonLab ...Thread below🧵https://t.co/q1ICykmVox