WEED rates

[Yanxiao Han]

Hi westpa developers,

I am trying to calculate the rate for my ligand dissociation. 

But the off rate is very low compared to the experimental value.

I tried to use the WEEDDriver to reweight, with the following line in my config file.

-----------------------------------------------------------------

  plugins:
    - plugin: westpa.westext.weed.WEEDDriver
      enabled: true
      do_reweighting: true
      window_size: 0.75
      bins:
        type: RectilinearBinMapper
        boundaries:
          - [0, 12.0, 40.0, 'inf']

---------------------------------------------------------------------

after reweight, I got the rates using the following codes:

----------------------------------------------------------------------

w_ipa -ao -W west.h5

w_red

red_fluxes = h5py.File(directh5file, 'r')['red_flux_evolution'][:]
red_rates = red_fluxes/(tau)

-----------------------------------------------------------------------

The new rate increases significantly with more iteration number but goes far bigger than the experimental value. And seems continue increasing as shown in the figure. Do you have any suggestions regarding this scenario? 

Also, I am not sure whether the way I obtain the rate is proper. 

I'd appreciate any of you suggestions.

Thank you,

Yanxiao

 

[Daniel Zuckerman]

Yanxiao, the developers can comment on whether they see a technical problem with your setup.  But I guess that you're also trying to do a calculation that is very challenging.  In some cases, even with WE simulation, certain observables may remain out of reach.

I guess you have done one WE run so far.  If you can possibly afford it, it's very useful to do at least one additional run.  Then you can compare estimates based on your raw WE data and the WEED analysis.  If those estimates are very different, then you know your system is challenging - that is, your WE sampling may be insufficient given the complexity of the system.

I should say that we are developing additional tools for rate-constant analysis.

You may also like to participate in WESTPA data club, which is open to anyone.  Details below.  —Dan

The data club is a virtual forum where you can get feedback at any stage of a WE project from Lillian Chong, myself, and members of our groups who are experienced with WE.

Presenters must obtain permission from their PIs, and we urge PIs to attend.  Besides the presenter and members of their group, only Drs. Chong, Zuckerman and members of their groups will be in attendance.

This is a ‘no strings attached’ offer.  There is no expectation for co-authorship for any participants.  Our interest is to see the best science done by WESTPA to keep our community strong.  The meetings will be 1:30 ET on 3rd Weds of the month, although alternative times may be possible.  Please contact me or Lillian if you are interested.

Thanks.  –Dan

---

From: westpa...@googlegroups.com <westpa...@googlegroups.com> on behalf of Yanxiao Han <hanya...@gmail.com>
Sent: Monday, January 22, 2024 1:09 PM
To: westpa...@googlegroups.com <westpa...@googlegroups.com>
Subject: [EXTERNAL] [westpa-users] WEED rates

 

--
You received this message because you are subscribed to the Google Groups "westpa-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to westpa-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/westpa-users/CAN7JJFN2%3DOaYDQcM31BVc%3D%2BGmvuPfX_zqS8WhZ%2BvxkQ_17ZvXQ%40mail.gmail.com.

[Lillian Chong]

Yanxiao, as Dan mentions, we do recommend in the best practices section of our WESTPA tutorials in LiveCoMS to carry out multiple WE runs if you can afford it. If you and your PI participate in the WE data club, it would be helpful to provide a plot of the time-evolution of your estimated rate (rate vs. WE iteration) without the application of the RED scheme.

Please note that the RED scheme is designed for only simulations that have been run with a "sink" state, i.e. trajectories that reach this state have been "recycled" to maintain a non-equilibrium steady state. Also, this scheme is not theoretically compatible with the use of WESS reweighting.

Best,

Lillian

To view this discussion on the web visit https://groups.google.com/d/msgid/westpa-users/CY5PR05MB9045913F852F3B7D6D7AA7EBB8742%40CY5PR05MB9045.namprd05.prod.outlook.com.

[Yanxiao Han]

Thank you so much Dan and Lillian for offering help.

I will talk to my PI about this.

Meanwhile, I will redo the analysis and make the plot.

It is definitely very interesting to obtain rates through the WE method.

Best,

Yanxiao

To view this discussion on the web visit https://groups.google.com/d/msgid/westpa-users/CAAHyvJtV%3DWWc8OH8bdwHs1QARL8ECffHWe-1macaKsax%3D0Jv-Q%40mail.gmail.com.