Posts on X

Xia Yang @YangLabUCLA at today’s #sbhd2019 introducing mergeomics, to combine multiple omics levels using genetic variation.

Want to leverage #scanpy's features for single-cell ATAC and DNA methylation? Finally, epiScanpy is out @biorxivpreprint - great collab led by Anna Danese from @colomemaria_epi lab. Please try&test at github.com/colomemaria/ep… and read at biorxiv.org/content/10.110… @bio_annaD

Come and work as PhD/Postdoc in #computationalbiology @HelmholtzMucEn and @TU_Muenchen, and join us in developing the next generation of #DeepLearning models for single cell genomics and spatial transcriptomics. bit.ly/2Qj0zVi #jobs #munich

EBI-colleagues around John @MarioniLab discuss how to understand biological noise in molecular recordings - great review: rdcu.be/bEfMw

Maybe I should suit up next time? 😂

@RegSprecherStS

Kanzlerin #Merkel diskutiert mit Studenten der @TU_Muenchen über Künstliche Intelligenz und besichtigt zusammen mit Ministerpräsident @Markus_Soeder an der Munich School of Robotics and Machine Intelligence herausragende Beispiele aktueller Forschung und Wissenschaft.

Quoted post

Chancellor Merkel is visiting the Munich School of Robotics and Machine Intelligence @TU_Muenchen - some impressive use cases of applied #AI shown!

Honest statement at a recent talk: ‘We googled, eh, did literature research and found ...’

Great contribution to #paperwriting I recently saw.

This is important and helps accelerate science - now let’s make data&code sharing an encouraged standard as well! #openscience

@obahcall

We encourage preprint sharing! Nature journals have supported & encouraged preprints for over 2 decades. Today we announce updated policy for all Springer Nature journals that includes info on preprint licensing, citation & communications with media. nature.com/articles/d4158…

Quoted post

Great initiative - often difficult to get funding for tools!

@biohub

Open source software is 🔑 to science. Many of the packages, libraries + applications crucial to biomedicine are built by researchers who volunteer their time + effort to make these tools available. We’re excited to announce support for #opensource tools bit.ly/2Q4G1zR

Quoted post

@segal_eran on algorithm-based personalized nutrition @scilifelab AI meeting. While the algorithm does not allow schnitzel, ice-cream was ok (for the person tested) if you also had hummus. Hopefully not at the same time.

Hiroaki Kitano, one of the founding fathers of #systemsbiology, talking @scilifelab AI meeting on limits of large-scale modeling. Manual annotation will not cut it, he recommends using AI to compile such maps, and jokes about an AI getting Nobel Prize in 2050

The #AI revolution has begun eating its own.

@kdnuggets

The end of average Data Scientist is near: @GoogleAI fully automated, End-to-End #AutoML #MachineLearning Solution for Tabular Data beats about 80% of human #DataScientists in recent @Kaggle competitions buff.ly/3073ZPB

Quoted post

Important piece @nature on AI regulation & ethics, and whether industry can fund research in this direction without corrupting it. Sounds like a classical conflict of interest to me. nature.com/articles/d4158…

Kauffman about one of the 'great scientific puzzles of our time', namely 'how life emerged from inorganic matter'. Interesting quite philosophical discussion!

@NaturePortfolio

.@Sara_Imari reviews Stuart Kauffman’s provocative take on emergence and evolution: ‘A World Beyond Physics: The Emergence and Evolution of Life’. go.nature.com/2VJGM6O

Quoted post

Welcome to @HelmholtzMucEn !

@caina89

I am very excited to announce will be joining the @PioneerCampus in beautiful Munich as a PI in October this year. I will work on genetics of #mentalhealth disorders, with a focus on how env #stress, #aging and variations in #metabolism affect disease risk and heterogeneity.

Quoted post

No this is @N_Rajewsky

@dagmarwaltemath

@fabian_theis @LifeTimeFET Are you the German national LifeTime partner to contact in step 1?

Quoted post

How to become an associated partner of @LifeTimeFET

@Alfons_Valencia @LifeTimeFET showing off the ‘most beautiful supercomputer’, from Barcelona. Pretty!

@LifeTimeFET meeting afterparty in the Berlin Hofbräuhaus - turns out it is not exactly Munich (570k away). 😂

Really glad you screenshotted my torus manifold picture - old-school math from my undergrad studies.

@NicoloCaporale

#LifeTimeOpening.Latent space to represent and interpret single cell data.@LifeTimeFET @MarioniLab @fabian_theis @jeanphi_vert @GoogleAI.AI advances and challenges for better medical predictions.Specific architectures for different data?Possibility of causal-effect actionability?

Quoted post

@LifeTimeFET’s goal in a nutshell. (@N_Rajewsky talking)

@GAlmouzni, @N_Rajewsky and now president Wiestler of @HelmholtzMucEn kicking off @LifeTimeFET - promising strong support of Helmholtz for this.

Interesting interview of Bertie Göttgens by Nicole Rusk @naturemethods - Bertie’s wish for a 2020 paper was mammalian genetic lineage tracing and computational prediction of future trajectories. Exciting, s.t. to discuss tomorrow @LifeTimeFET! go.nature.com/2Sd79jQ

There’s many reviews for DL in the life sciences such as nature.com/articles/d4158… or our genomics-focussed recent one rdcu.be/bvNef - for more general ML you can essentially look at any paper from computational biology I would say.

@ranjeet_sngh

@machine_ml @machinelearn_d @gtesta72 @LifeTimeFET @fabian_theis @HelmholtzMucEn Could you pls redirect me to appropriate resources

Quoted post

Looking forward to the @LifeTimeFET meeting tomorrow - excited to discuss how we can leverage single-cell techniques and machine learning to deal with clinical problems. Interesting lineup of speakers for sure, here’s the livestream: lifetime-fetflagship.eu/index.php/life…

Exactly - kind of nice since this means we won’t be out of job that quickly

@ivivek87

@fabian_theis in consensus with the belief that there is no perfect data in real world, I would be not worried that we will not have the "right" ML model. It is totally data-driven and problem in hand.

Quoted post

Hate to break it but there is no ‘right’ machine learning model - en.m.wikipedia.org/wiki/No_free_l…

@COMPUTERWOCHE

[Anzeige] "Trial & Error" ist keine Option.

Quoted post

😂 paper figure Easter eggs, that was a novelty for me.

@ksuhre

@fabian_theis @NPRougier @PLOSCompBiol Rule 11: show surprising and interesting details medium.com/datadriveninve…

Quoted post

You make it sound all bad, but I think we scientists need to be able to communicate our ideas clearly (without oversimplification, agreed!), and this includes good figures.

@chris_penfold

This is unfortunately true 😞 Look and politics are queen/king. The march of the style-over-substance, slick, sleek, Nature-esque publishing of modern science.

Quoted post