search

prody / ProDy

A Python Package for Protein Dynamics Analysis
http://prody.csb.pitt.edu
Other
417 stars 153 forks source link

Building ANM for Cryo-EM map #1113

Closed gmk19001 closed 4 years ago

gmk19001 commented 4 years ago

I am trying to build an ANM model for a cryo-em map using the tutorial listed on the ProDy website. However, after I parse the structure using parseEMD and create the ANM instance, I cannot build the hessian matrix. The error listed says: "coords must be a Numpy array or an object with getCoords method"

I then try to convert my emd to numpy array and receive errors.

Is there a step that I am missing or something that I can do to resolve these errors?

jamesmkrieger commented 4 years ago

As it says, you need an object with the getCoords method or a coordinate array to build a Hessian matrix. If you use parseEMD with the appropriate parameters then you should get an AtomGroup, which has that method and contains a set of pseudoatoms that have been fit to the cryo-EM density map using the topology representing network algorithm (using the TRNET class).

However, that function by default gives you the EM density map (an EMDMAP object) directly. If you were to try to build a Hessian matrix with that, then it would not work and likely give you the error you get.

Can you check what your output is and what ProDy version you're using please? Also, did you use the same command as on the website tutorial?

Many thanks James

jamesmkrieger commented 4 years ago

It looks like the same behaviour of building coordinates using TRNET is triggered by the parseEMD command on the website tutorial with both the latest PyPI release (version 1.10.11 from 25th October 2019) and the latest development version on Github (which has been version 1.11 for a long time). I am still waiting for it to run to see whether the output is correct but can you update your ProDy to one of those and see if that fixes the problem for you?

Many thanks James

jamesmkrieger commented 4 years ago

I definitely get an AtomGroup with both ProDy versions and the Hessian gets built fine too.

gmk19001 commented 4 years ago

Hi James,

Thank you for getting back to me and for this useful information. I think that it was my version of that was giving me trouble. I was using 1.10.10. I reinstalled and everything is working well now. All the best, Genny

Genevieve Kunkel Graduate Student, Tarakanova Lab Dept. of Mechanical Engineering, University of Connecticut

genevieve.kunkel@uconn.edu (mailto:genevieve.kunkel@uconn.edu)

On Aug 10 2020, at 2:02 pm, James Krieger notifications@github.com wrote:

Message sent from a system outside of UConn.

I definitely get an AtomGroup with both ProDy versions and the Hessian gets built fine too. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub (https://github.com/prody/ProDy/issues/1113#issuecomment-671502773), or unsubscribe (https://github.com/notifications/unsubscribe-auth/AQRROSTBROPOOIZKKIO5YD3SAAY33ANCNFSM4PY6O2LQ).

jamesmkrieger commented 4 years ago

Hi Genny,

You're welcome. I'm glad to hear it works now.

Best wishes James