Jeff van Santen

Scientist and developer, outdoors enthusiast, bike + beer lover. Also a big nerd. I work on @NatProdAtlas
🌐

[Marnix Medema @marnixmedema]

Online @NatRevDrugDisc : our review on artificial intelligence for natural product drug discovery! A collaborative product emerging from a @lorentzcenter workshop, co-chaired by @GJPvWesten @HirschLab #RogerLinington @serina_robinson and myself. https://t.co/QKqHrMEAGp 1/n

[Jeff van Santen @JeffAvS]

@georgeprbenson I’m getting my boost today too!

[Nadja Cech @nadjacech]

Excited to have NP Analyst, developed by the Linington Group @NatProdAtlas as part of the #naturalproducts research toolbox.

This tool predicts active constituents from MS #metabolomics datasets

Congrats to the authors and thank you!
@JeffAvS
https://t.co/KndTbQLcLj

[Jeff van Santen @JeffAvS]

Proud to see the latest from my with the @NatProdAtlas / the Linington Lab at @SFUChemistry. Although it's not longer my FT, I am thankful for all the excellent collaborations I was able to participate in while there.

[NP-MRD @np_mrd]

Appearing soon in the Nucleic Acids Research 2022 Database Issue:
NP-MRD: the Natural Products Magnetic Resonance Database https://t.co/IpLZvHkhGY

[Jeff van Santen @JeffAvS]

@__Joe_Egan__ Talking about RAM? If you’re using WSL that’s probably the culprit

[Jeff van Santen @JeffAvS]

@ChasingMicrobes I’ll have to learn the “RStudio algorithm”!

[Jeff van Santen @JeffAvS]

Incredibly proud of the latest work with the @NatProdAtlas. Thanks to the great team of collaborators who contributed to making #openscience better for everyone!

[J. Am. Chem. Soc. @J_A_C_S]

Bimodal Evans–Polanyi Relationships in Hydrogen Atom Transfer from C(sp3)–H Bonds to the Cumyloxyl Radical. A Combined Time-Resolved Kinetic and Computational Study https://t.co/Hrp5VHQanW

[Jeff van Santen @JeffAvS]

@MWMullowney Oh man, we sure did eat a lot of late night pizza... miss hanging out with cool scientists!

[Jeff van Santen @JeffAvS]

@BlakeAntcliff @LinusTech @msiUSA Not really sure that's true given link deets are part of the video/challenge...

[Justin van der Hooft @jjjvanderhooft]

The Paired Omics Data Platform just got out in @nchembio! Excited & Proud!😎 Platform: https://t.co/MJyJAxqIZA; Paper: https://t.co/IEQIuI53yT; or as direct read: https://t.co/7FkdB4tNbb A little thread!
#genomics #metabolomics #collaborative #community #ProudPI @WU_BioInfo @WUR

[Jeff van Santen @JeffAvS]

@NadineZiemert @NatProdAtlas @ZENODO_ORG JSON files now available for download @ https://t.co/FGVVEalfea

[Jeff van Santen @JeffAvS]

@NadineZiemert Just to follow up on this: For the latest @NatProdAtlas which has 29,006 total (11,264 bacterial) compounds there are 466 total (313 bacterial) unique classes predicted by NP Classifier. I'll upload the results dataset to @ZENODO_ORG shortly along with ClassyFire ones.

[Jeff van Santen @JeffAvS]

@thinkgenome @NadineZiemert @NatProdAtlas @ZENODO_ORG Primary metabolites may also be a missing category.

[Jeff van Santen @JeffAvS]

@thinkgenome @NadineZiemert @NatProdAtlas @ZENODO_ORG Well @NatProdAtlas is a microbial NP database, so the remaining classes must be from other realms of life

[Jeff van Santen @JeffAvS]

@jjjvanderhooft @NadineZiemert @NatProdAtlas @ZENODO_ORG I think we may even be there already on our dev DB! We're overdue for a release, but things have been busy as ever

[Jeff van Santen @JeffAvS]

@jjjvanderhooft @NadineZiemert @NatProdAtlas @ZENODO_ORG I'm already thinking about how to integrate these/Classyfire into the NP Atlas webpage effectively 🤓 Adding them directly the the flat file download is an option, but for now they'll just be in a separate JSON organized by NPAID as key -> results object for each access.

[Jeff van Santen @JeffAvS]

@mingxunwang @NadineZiemert If you move to POST then you might also have other problems like not being able to embed as an image, etc. I don't honestly have the perfect solution

[Jeff van Santen @JeffAvS]

@mingxunwang @NadineZiemert I might be thinking of an instance where I was using FastAPI instead of Flask, so you might be right about auto decoding. Base64 seems like a decent solution (although still not URL safe without encoding b/c '+' '/' + '=').