David Feng

Director of Scientific Computing, Allen Institute for Neural Dynamics

[Lydia Ng @2queues]

Sharing🤯video to get you through Friday!
Credit: @E_cross_B

@Blender animation of the Allen Mouse Common Coordinate Framework (#CCF). Outer structures "explodes" out to reveal inner brain structures.

@AllenInstitute

[Josh Moore @notjustmoore]

Out now! "#OMEZarr: a #CloudOptimized #BioImaging #FileFormat with Int'l #Community Support" 😍

Huge thanks to the 68 co-authors, across 32 institutions (see the quoted tweet) 🙌

If you use #OMEZarr & weren't involved, let's start the next report! 📝

https://t.co/iSFOr3RYYU

[David Feng @dyfbrain]

@adam_k_glaser Congrats @adam_k_glaser and team! Some of the most beautiful (and largest!) data I have ever seen 🤩

[Adam Glaser @adam_k_glaser]

How do you image the large and the small at the same time? We developed new 🔬 technology to image centimeter-scale specimens - including whole mouse brains 🧠 - with diffraction-limited resolution and without sectioning. #mesoscale #imaging

https://t.co/Kl7ZURNpdM

🧵 (1/n)

[David Feng @dyfbrain]

@BWJones @grimmrad @RetoPaul We usually do one Zarr "file" per acquisition session -tend not to need to update/rebuild anything. Although I feel like I may be misunderstanding.

[Alessio Buccino @buccino_alessio]

Do you use #neuropixels or #highdensity probes? Are your recordings filling up your hard drives?
We got you covered!

In the first preprint from @AllenInstitute for Neural Dynamics, we looked at ways to reduce the footprint of #ephys data.

https://t.co/sH1Lt5k3Xw

(1/n)

[David Feng @dyfbrain]

@buccino_alessio @AllenInstitute Congrats @buccino_alessio and team!

Effective data compression is often treated as an afterthought, but long term it saves everyone time and $$$. Hopefully everyone else can take advantage of this as well!

[David Feng @dyfbrain]

@grimmrad @BWJones @RetoPaul We like OME-Zarr for this. Zarr comes with a whole bunch of lossless codecs that are trivial to apply. For our SPIM data we routinely get 5-8X lossless compression.

[Josh Moore @notjustmoore]

"Toward #scalable reuse of #vEM data: #OMEZarr to the rescue" just hit the shelves!📚

https://t.co/kOJBPfw54o

Special thanks to co-authors @normanrz & @BogovicJohn, co-artist @DrHenningFalk, and volume (ha!) editors @EM_STP @KedarNarayan2 & @PaulVerkade

[Josh Moore @notjustmoore]

Latest report from the #OMENGFF #Community describing #OMEZARR, a #CloudOptimized #BioImaging #DataFormat, is out now on #bioRxiv!

If you have issues with big, slow, or otherwise unwieldy imaging data, reach out or take a look for a #FAIR-er option.

https://t.co/OfSZmiMZuA

[David Feng @dyfbrain]

@_martinthoma https://t.co/XDtBHm2xHF

[David Feng @dyfbrain]

@neuromusic ding ding

[David Feng @dyfbrain]

I just learned the phrase "screaming snake case" and can't stop giggling.

[David Feng @dyfbrain]

@chrhagel As with most things, it's funnier without the explanation: https://t.co/K3rrF2Lbd6

[David Feng @dyfbrain]

@neuromusic @jedsundwall https://t.co/SNsdxVAq89

[David Feng @dyfbrain]

The CCF is on Mailbox Peak now. @2queues @neuromusic

[David Feng @dyfbrain]

@adam_k_glaser @AllenInstitute @MesoSpim Maybe next time I will keep my big mouth shut when I tell an acquisition team to "just collect better data." This data is huge!

[Patrick Mineault @patrickmineault]

What do scientists here use to keep track of experiments, subjects, local datasets, etc? Spreadsheets, Airtable, structured SQL databases? MongoDB or Firebase? Some special software because of PII or HIPAA? What has worked for you in the past, and what hasn't?

[David Feng @dyfbrain]

Quote of the day (~5 years ago): "You're just secretly trying to convert all of our scientists to software engineers, aren't you?"

Here's to hoping it's now safe to set aside my secret devops agenda.

[David Feng @dyfbrain]

Quote of the day: "Just a quick heads up - I am going hunt you all down. For metadata."

[David Feng @dyfbrain]

Surrealist pokemon, so fun.

[David Feng @dyfbrain]

More adventures in generative models. I threw pokemon at https://t.co/CDUeuKh8rl without touching anything and the results are fun.

[David Feng @dyfbrain]

I switched the loss function from L1 to MSE and the style change is cool.