Dylan SERILLON

[Dylan SERILLON @DSerillon]

@AgnieszkaSzumna Hi Agnieska!

For the sampling (molecular dynamic) we are using explicit solvation model.
For the calculation at semi empirical level using GFN2B gbsa implicit solvation model is used.
For both multiple solvation model are parametrised (water, dichloromethan, chloroform...)

[Dylan SERILLON @DSerillon]

@MichaelDWard7 Hi Michael, extracted from binding DB + SAMPL challenges, I have almost 570 free binding energy (most of them are ITC measurement) interacting with 27 different host system.

[Dylan SERILLON @DSerillon]

@CMozaceanu @NOAH_ITN @AgSchalley Amazing poster Cristina!

[Dylan SERILLON @DSerillon]

@ralvarezyebra @NOAH_ITN Interesting poster & Molecular Dynamic !

Why do you decide to use AM1-bcc charge instead of RESP charge that are generally more accurate?

As well, do you compared the dynamic of the system without guest molecule ?

[Dylan SERILLON @DSerillon]

@abarbieri50 The main differences between both methods is the association time. With SaMD it sometimes take longer to associate.
In what we have learnt, SaMD is a good option for very flexible system, but doesn't work well on the rigid system that doesn't favorised the inclusion process.

[Dylan SERILLON @DSerillon]

@abarbieri50 Well concerning the simulation time, there is nothing comparable between Machine learning and Molecular dynamic.

ML take like 15 second to give a results.
Concerning the differences between MD and SaMD if we compared for 500ns simulations, there is ~1 hour difference.