This code is for determining the chemical shifts of exchanging states from CEST experiments using DNN To use the code, the file containing the weights for the trained network must be downloaded. This is available here. Once downloaded place the weights in a directory and point to that directory when running the programme. The code for creating untrained models is contained in the models folder.
The script can then be run as follows:
./DNNautoCEST.py
-sfrq SpecFreq
-xcar CarrierFreq
--datadir DATADIR
--expname EXPNAME
-model Directory_with_weights
The out, conf, gpu arguments are optional and will default to 'result', 0.4, and -1
The script has been written and tested with the above dependencies. Performance with other module versions has not been tested.
Please cite the following article if you use this script in your work:
Karunanithy G., Yuwen T., Kay LE., Hansen DF. Towards autonomous analysis of Chemical Exchange Saturation Transfer experiments using Deep Neural Networks https://doi.org/10.26434/chemrxiv-2021-r1cmw