Govinda Kamath

Opinions my own.

[Govinda Kamath @GoviKamath]

@anshulkundaje @fabian_theis To be honest, I cant see how the study concludes anything meaningful.

If you sample 1% of 22M cells from 249 datasets; you'd be left with a sketch that has representatives of all the datasets. If you want to test scaling laws; sampling has to be at a dataset level.

[Steven Wu @zstevenwu]

There's growing evidence that LLMs can p-hack. That should worry us.

But p-hacking also points to something bigger: a data science multiverse of defensible analytical choices. LLMs make it cheap to search that space for favorable results.

In our paper, joint work with excellent collaborators: Martin Bertran and Riccardo Fogliato, we build a pipeline that uses LLM agents to systematically map this multiverse. We believe our experiments could inform what future scientific transparency should look like. 🧵

[Pall Melsted @pmelsted]

Excited to share this preprint that describes my latest work on using GPUs to accelerate processing of RNA-seq data.

The title says it all: "RNA-seq analysis in seconds using GPUs" now on biorxiv https://t.co/2JrOfsxNFV

Figure 1 shows they key result

[Harveen Singh Chadha @HarveenChadha]

Just to confirm all Sarvam models are foundational and trained from scratch… our models are not finetuned of any open source models

[Hemant Mohapatra @MohapatraHemant]

"And you shouldn’t try — it’s your job to try anyway.”

105B. Pre-trained from scratch. @SarvamAI .

[Stephen Williams @SvenAT]

I'm excited to to announce our preprint describing automated cell annotation for human and mouse in @10xGenomics cellranger 9.0+ in collaboration with the @broadinstitute! https://t.co/60gae6fiRy

[bioRxiv Bioinfo @biorxiv_bioinfo]

Accelerating scRNA-seq analysis: automated cell type annotation using representation learning and vector search https://t.co/z3QHlPNVhz #biorxiv_bioinfo

[Govinda Kamath @GoviKamath]

@Kangwook_Lee Really cool idea!! May use it on something I need to do some cleanup on.

[10x Genomics @10xGenomics]

📢 New Feature Alert 🎉 The beta version of Automated Cell Annotation is live in Cloud Analysis! Reduce the burden of manual annotation & visualize cell types instantly to get more accurate #singlecell insights faster. Try it now: https://t.co/77dztDXhBT

[Govinda Kamath @GoviKamath]

@fiandola Is the Demo broken? I was trying to run it on the example images, but it seems to error out.

[Govinda Kamath @GoviKamath]

@prateekgv @Rajesh23MD Brave is great -- especially on mobile. The only mobile browser with adblock AFAIK.

[bioRxiv Bioinfo @biorxiv_bioinfo]

ntsm: an alignment-free, ultra low coverage, sequencing technology agnostic, intraspecies sample comparison tool for sample swap ... https://t.co/lny2YAvpau #biorxiv_bioinfo

[Govinda Kamath @GoviKamath]

@prateekgv This word-salad is just missing some reference to "blockchain-technology"... 😂😂😂

[Joey Arthur @JoeyArthurOmics]

Stoked to see the software I worked on during my PhD run on 4,000+ human genome samples - big thanks to my co-authors (and especially Bo Zhou) for leading the charge

https://t.co/o3aXTXPBqy

Methodologically, I especially like the coming together of bioinformatics algos, [1/2]

[bioRxiv @biorxivpreprint]

High resolution mapping of the breast cancer tumor microenvironment using integrated single cell, spatial and in situ analysis of FFPE tissue https://t.co/70zWg8oZUn #bioRxiv

[Alexis Vandenbon @alexisvdb]

Meanwhile, the single cell field

[Govinda Kamath @GoviKamath]

@skbutp Yes.

[Govinda Kamath @GoviKamath]

@skbutp Shubhakrit is the name of this year. Vikram is the name of the calendar (of which this is 2079th year)

[Govinda Kamath @GoviKamath]

@skbutp The American founders he's referring to considered black people as 3/5th human and native Americans as non-human.

And documents they wrote are trotted out as some sort of champion of "human rights".

[Govinda Kamath @GoviKamath]

@skbutp Paa, Srilanka was called Ceylon till 1970. Stalin died in 1953.