Anton Siomchen

Computational chemistry and chemical task automation |
AL/ML | Jagiellonian University | Ryvu Therapeutics

[Anton Siomchen @vctr_machine]

@uwteam @JanuszT54799155 Miałem ten problem, chyba po tym jak wybierzesz informacje o twojej paczce na końcu jest opcja człowieka (to 99% nie pomaga ale można spróbować)

[Anton Siomchen @vctr_machine]

@zavsievich Не, ну это, конечно, возмутительно, срочно топаю говорить, что пора уходить из этого гитлаба (у них ведь даже lab в названии есть, как так?)

[Anton Siomchen @vctr_machine]

@andrewwhite01 And then, when you think, that you are done, you are spending 20% fixing CUDA for pytorch

[Martin Borch Jensen @MartinBJensen]

To clarify, I'm not saying there's no place for AI in discovery. But in that context, better targets/compounds >> "making discovery faster and cheaper".

Specific pitch "it costs $B to make a drug, my molecule generator can make a thousand drugs by breakfast" = 🙈

[Anton Siomchen @vctr_machine]

@MartinBJensen It is funny, but a couple of months ago, I was trying to convey the same message with this exact image, while talking about AI powered vHTS

[Anton Siomchen @vctr_machine]

@AnthonyC1Nash Is it really that bad? Maybe migration from conda to mamba would help you?

[Anton Siomchen @vctr_machine]

@HarmonicDiscov Wow, this is interesting

[Anton Siomchen @vctr_machine]

@zavsievich Носки - топ, а почему коала на борде?

[Anton Siomchen @vctr_machine]

It was supposed to be small function to check if the molecule is suitable for analysis, but #copilot had more pessimistic option...

#compchem #rdkit #smiles

[Anton Siomchen @vctr_machine]

@AnthonyC1Nash https://t.co/FetCOeaCni
Works pretty well for couple residues, don't know is it suitable for bigger gaps

[Anton Siomchen @vctr_machine]

@leecronin Whole publishing system should be built from scratch. Situation, when results of tax-funded research are placed under paywall is really sad (not to mention, that in developing countries it is even bigger problem - universities cannot afford such high subscription costs)

[Anton Siomchen @vctr_machine]

@ML_Chem Super interesting, it's even might help me with research for master thesis

[Anton Siomchen @vctr_machine]

@mooonradio А в Польше ушла эпоха: в North Fish нет больше супов по 4.49

[Anton Siomchen @vctr_machine]

@andrewwhite01 It works, but after submitting smiles appears constant loading ring instead of buttons text, is it supposed to do something else except drawing a molecule?

[Anton Siomchen @vctr_machine]

@AnthonyC1Nash Also check MODELLER by Andrej Sali it might help with homology modelling of structure

[Anton Siomchen @vctr_machine]

@MSiggel @AnthonyC1Nash Alternatively Swissmodel might be helpful to predict secondary structure, if there are any similar proteins available

[Anton Siomchen @vctr_machine]

@PaulRobustelli Thank you, will check it out, both books are even available at my uni library

[Anton Siomchen @vctr_machine]

Hi #compchem twitter, I would like to say thank you to @PaulRobustelli for his free CHEM101.6 course. I want to make some MD simulations for my master's thesis, and this course has shown me how flexible #openmm is

Below 10 ns simulation for T4 lysozyme and its ligand

[Anton Siomchen @vctr_machine]

When you think that it might be a good idea to start grad school next autumn, GROMACS reminds you:

#compchem #AcademicTwitter #PhD #gromacs

[Anton Siomchen @vctr_machine]

@jchodera @UCSF It's seems to be international problem, probably except Swithland and Netherlands, that PhD stipend is below living wage, it's really sad

[Anton Siomchen @vctr_machine]

@anunesalves @PaulRobustelli Still a lot work to do, a lot concept to learn, but, as for partially new field for me, it goes well