water leg broken

[xy21hb]

Dear all,

I have got error like below. Quite different from previous common errors, I only have errors from the water system leg,

For each transformation, e.g. FXR_12 ---> FXR_88, I go to /PyAutoFEP-master/docs/tutorial01/tutorial1/FXR_12-FXR88/water/md, l can have logs like below, I checked the feprun... log and rerenjoblog....log, there is no notable error msg inside,

However, there is nothing in /analysis, there seems to be no traj or analysis generated. Moreover, in /rerun, I can have files like,

However, in the counterpart of protein leg, everything seems to be okay,

So why does the water leg only have problems, but not protein leg? Thanks!

[luancarvalhomartins]

From your first image, it seems the water leg did not run. Would you mind checking the gmx mdrun logs for the water leg?

[xy21hb]

mdout_nocont.log mdout.log

Above are the md files, e.g. under /PyAutoFEP-master/docs/tutorial01/tutorial1/FXR_12-FXR_88/water/lambda11/md.

Below are lambda files, e.g. under PyAutoFEP-master/docs/tutorial01/tutorial1/FXR_12-FXR_88/water/md/lambda11, lambda.log

I also noticed in this directory, plumed.dat has nothing inside.

[luancarvalhomartins]

Thank you for the detailed information.

The lambda.log shows you are running only 0.6 ps (nsteps = 300, dt = 0.002). Because nstxout = 1000, you only get 1 frame. PyAutoFEP was able to rerun and collect data even using a single frame! On a second thought, I think everything worked (to the extent it could be possible). If the error you are concerned about is the tar one, that's harmless.

What are you trying to do, exactly?

[xy21hb]

Dear Luan,

Many thanks for your reply.

Yes, I understand that the windows are short and meaningless to get statistics. All I want to do is just to test the workflow on a single non-cluster local desktop using serial run. Since regular calculations takes much time and it is hard to get error message if there are errors in the middle of the runs.

I extended the simulation length to longer times below. Still it has the same problem, i.e. the water leg do not have anything under analysis. I wonder why only water leg has this problem specifically, but not the protein-ligand complex leg, no matter how small the lambda window is.

[cid:4ce6bc42-f5f9-4088-8837-bb74535c6638]

However, it seems to be working for analyze_results.py,

[cid:fdd94285-7c5d-4ff6-9897-636b5dfb534e]

Of course, those ddG might be meaningless due to short windows.

Best, Yao

---

From: Luan Carvalho Martins @.> Sent: Wednesday, July 26, 2023 10:59 PM To: luancarvalhomartins/PyAutoFEP @.> Cc: xy21hb @.>; Author @.> Subject: Re: [luancarvalhomartins/PyAutoFEP] water leg broken (Issue #101)

Thank you for the detailed information.

The lambda.log shows you are running only 0.6 ps (nsteps = 300, dt = 0.002). Because nstxout = 1000, you only get 1 frame. PyAutoFEP was able to rerun and collect data even using a single frame! On a second thought, I think everything worked (to the extent it could be possible). If the error you are concerned about is the tar one, that's harmless.

What are you trying to do, exactly?

— Reply to this email directly, view it on GitHubhttps://github.com/luancarvalhomartins/PyAutoFEP/issues/101#issuecomment-1652486462, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ADM4TWQCFVJZHGL6G5D5WRDXSGAKZANCNFSM6AAAAAA2MV5RAQ. You are receiving this because you authored the thread.Message ID: @.***>

[xy21hb]

[luancarvalhomartins]

Currently, no analysis is performed for water leg. I hope I find time to add that in the future.

From your last message, it seems everything seems to be working fine.