best-curve001.txt file in EOM and frequency of structures

Linear (OLIGOMER), and non-linear (MIXTURE) analysis, singular value decomposition (SVDPLOT), addition of missing fragments (BUNCH, CORAL), analysis of flexible systems (EOM/RANCH & GAJOE), flexible refinement of high-resolution models (SREFLEX)
Post Reply
Message
Author
vb
Member
Posts: 3
Joined: 2015.11.27 13:37

best-curve001.txt file in EOM and frequency of structures

#1 Post by vb » 2015.11.27 14:01

I used to run a version of Gajoe, a few years ago, that provided the output file best-curve001.txt. This output file was very interesting because it contained the list of each 10,000 structures with their respective frequency in the final ensemble, and this allowed us to analyse differently the selected ensemble of conformers (inter-domain distance distribution, for example, etc....).
I again need to get this information from my EOM simulations, but this file is not provided anymore in the new versions of EOM/Gajoe. How can I get it ? Is there any option to type when running EOM/Gajoe that would allow us to get it ? Or what version of EOM is the latest with this option, and which I could thus use ?

Many thanks !
Attachments
best_curve002.txt
(12.88 KiB) Downloaded 198 times

User avatar
Hayds
Active member
Posts: 101
Joined: 2008.05.21 19:01
Location: EMBL, Hamburg

Re: best-curve001.txt file in EOM and frequency of structure

#2 Post by Hayds » 2016.01.12 10:39

Dear vb,

We did indeed remove this file at some point and as you noted it is no longer part of the output in EOM 2.0. I was not sure people actually used it. I will see what can be done to recreate it (or reintroduce it into EOM again).

Cheers,

Haydyn

vb
Member
Posts: 3
Joined: 2015.11.27 13:37

Re: best-curve001.txt file in EOM and frequency of structure

#3 Post by vb » 2016.01.12 11:34

Dear Haydyn,
Thank you very much. Yes, if you could reintroduce it somehow, it would be great. This would avoid me to use old versions of EOM, as the results are not always completely reproducible with the newest version. Please let me know as soon as I can use the latest version of EOM providing this file.
Many thanks,
vb

User avatar
Hayds
Active member
Posts: 101
Joined: 2008.05.21 19:01
Location: EMBL, Hamburg

Re: best-curve001.txt file in EOM and frequency of structure

#4 Post by Hayds » 2016.01.22 11:46

Dear vb,

it will take a bit of time to reimplement a best_curve001.txt output file from EOM 2.1. In the meantime you can use the -w flag to write out all the work files used by the genetic algorithm. The GA_CYCLE files list the selected curves for each ensemble, from these files you could estimate the frequency of selection of each curve in the entire GAJOE run.

eg. on the command line:

gajoe -i juneom.int -w -t 1 data.dat

this will generate a series of files (GA_CYCLE_1.txt ... etc) with the following format:

Gener Best
1 10) chi^2: 2.97 - ensemble: -3-18-31-42-55-77-77-77-77-82-87-
1 9) chi^2: 2.90 - ensemble: -2-19-21-49-83-90-
1 8) chi^2: 2.90 - ensemble: -2-18-18-19-31-31-31-44-46-48-67-72-88-98-
1 7) chi^2: 2.88 - ensemble: -6-8-11-17-19-21-38-44-46-51-52-55-58-59-70-71-81-87-
....

You should be able to extract the number of times each curve is selected from these files and determine a relative frequency.

Hope it helps!

Cheers,

Haydyn

vb
Member
Posts: 3
Joined: 2015.11.27 13:37

Re: best-curve001.txt file in EOM and frequency of structure

#5 Post by vb » 2016.01.22 17:34

Dear Haydyn,

thank you very much. It worked ! But now, which structures are to be taken for the final Rg/Dmax distribution ? To what I understand, this Ga_Cycle_1.txt provides the number of each structures in the 40% best ensembles (a liste of 20 ensembles are provided in this file when 50 ensembles, as default parameter, are used against the experimental data), at each generation. Are all the structures of all the 20 ensembles of all the 1000 generations, used for the Rg/Dmax distribution ? Or only the structures of the 20 ensembles of the best (i.e. the last) generation?
Thank you again.

vb

Post Reply