Lorenzo Gaifas

PhD student doing Cryo-ET 🔬 Computational Methods nerd 🖥️ Core dev at @napari_imaging
#OpenSourceAcceleratesScience

[Lorenzo Gaifas @brisvag]

FYI my twitter account was compromised. I'm back in, but if in the last few days weird things happened, that's why.

[Lorenzo Gaifas @brisvag]

How embarassing (yet unsurprising), a decision like this coming from our head of state. And everyone doing backflips to prove they are "respectful to the dead". He didn't studdenly stop being a piece of shit just because he's dead.

[Lorenzo Gaifas @brisvag]

One day of national mourning (and 7 days of holidays for the parliament??) for a man who - for my whole existence - embezzled, dealt with the mafia, abused women, evaded taxes, was BFF with the worst dictators (Putin, Gaddafi), constantly embarassed us to the eyes of the world...

[GRAL (Structural & Cell Biology in Grenoble) @LabexGRAL]

@TimminsJoanna and Irina Gutsche from @IBS_Grenoble present collaborative work on cell division of radioresistant Deinococcus radiodurans 🦠, including PhD work of @brisvag funded by @LabexGRAL

[Lorenzo Gaifas @brisvag]

@SjorsScheres @a_punjani @structurabio @The_MRC I personally find that reasonable. In fact, not being able to distribute a GPLv2 program without any restriction to the GPLv2 freedoms does seem to go against GPLv2 to my non-legal eyes...

[Lorenzo Gaifas @brisvag]

@a_punjani @SjorsScheres @structurabio @The_MRC It would be interesting to hear some details on the perspective of @The_MRC legal team on this; it's always hard as a layman to understand implications of licenses, and the patents section in GPLv2 is lacking compared to v3 (which exists partly for this reason) @SjorsScheres

[Lorenzo Gaifas @brisvag]

@a_punjani @SjorsScheres @structurabio This seems to imply that @structurabio's patent for academia is 100% compatible with GPL(v2) (and that this was confirmed/discussed with the @The_MRC legal team) and therefore it was unnecessary to remove the code from #RELION to comply. Can you confirm this is what you mean?

[Lorenzo Gaifas @brisvag]

@Czaki_PL @liu_hanjin 😱

[Lorenzo Gaifas @brisvag]

@Czaki_PL @liu_hanjin Wow! I want this for layer names in napari!

[Lorenzo Gaifas @brisvag]

@DTegunov Neat! And if you're copying and renaming mrcs because you're using https://t.co/TPAkmh0oI0, you might like my upgraded version of the script that does that (and more) for you :) https://t.co/HBnrAYPDCS

[Lorenzo Gaifas @brisvag]

@jnuneziglesias @jaime_rgp @github There's also commenter:jni for that :)

[Alister Burt @AlisterBurt]

A bit 😴 after an amazing week in the bay area for a @cziscience workshop on the computational frontiers of cryo-ET
Here are my slides on @napari_imaging for #teamtomo
Original @ Google slides: https://t.co/1LWMXS2P2w
Zenodo archive: https://t.co/fOymTY6Q9G

[Alister Burt @AlisterBurt]

live weighted backprojection demo in @napari_imaging 🙂

[Lorenzo Gaifas @brisvag]

@liu_hanjin Yes! I though about trying this before with Textual, but it's a lot of work :p

[Lorenzo Gaifas @brisvag]

@Czaki_PL @napari_imaging @webknossos @jostriebel @EuroSciPy Yup, agree!

[Lorenzo Gaifas @brisvag]

Since there was some interest at #tomo2020: I made a simple @napari_imaging plugin to quickly interpolate N-dimensional label across a single dimension: https://t.co/oe2Q37VOxd. So you can stuff like you did in Amira, but with FOSS ;)

[Lorenzo Gaifas @brisvag]

@Czaki_PL @napari_imaging I used a SDF-based approach that I learned from @webknossos, thanks to @jostriebel at @EuroSciPy. It's a simple weighted average over signed distance fields, but it works like a charm!

[Lorenzo Gaifas @brisvag]

@ChailletMarten @napari_imaging If you just want an mrc, mrcfile is great :) (blik should take care of this soon as well)

[Lorenzo Gaifas @brisvag]

@ChailletMarten @napari_imaging I don't know what formats are typically used for this in our field, so I haven't done any writer for it yet. But you can access the data of the layer which is a simple numpy array, so if you know how it's easy for you to do!

[Lorenzo Gaifas @brisvag]

@s_e_taylor @napari_imaging I confirm it works, at least in terms of math. I'm not entirely sure about the cutoff used for the distance field, at first glace it seems a bit low (i.e: labels "fade away" too fast between slices). Let me know how it performs on real data!

[Lorenzo Gaifas @brisvag]

@s_e_taylor @napari_imaging In *theory* yes, but please let me know how wrong I am :p