 |
|
AMS 3.0 - predicting post translational modification sites
Material
Paperhttp://www.biomedcentral.com/1471-2105/11/210/abstract OpenMS docu:
http://www-bs2.informatik.uni-tuebingen.de/services/OpenMS/search/OpenMS/doc/html2/classes.html Programming with OpenMS (ie use OpenMS classes in your own C++ project):
http://www-bs2.informatik.uni-tuebingen.de/services/OpenMS/search/OpenMS/doc/html2/external_code.html You can probably avoid using 'my_custom_lib', as one main program should suffice.
Code
C code available from paper (Suppl. Material):
-
http://code.google.com/p/automotifserver/downloads/list - Download Predictor_source.zip (done)
→ download Phospho_PKA Training.zip for training a multilayer perceptron (MLP) yourself (done)
→ download AMS3_Consensus_distributable_v1.tar.gz for already trained MLPs to test expression patterns yourself (done)
- here you can find the phospho-PKA model as well as other post-translational modifications (PTMs)
- --> C++ code for prediction engine (todo) - either wrap C code or rewrite
→ original source code as an external project in OpenMS: (done)
→ you can use the attachment (https://svn.imp.fu-berlin.de/seqan/trunk/teaching/master/P4-2010/Proteomics/coding-projects/PTMs/Predictor.tar.gz) and then follow these instructions:- change the path to OpenMS in line 19 in CMakeLists.txt and type "cmake ." and then "make" (ignore the warnings)
- follow the readme-instructions (http://code.google.com/p/automotifserver/wiki/Predictor_1_3_ReadMe) to train a model using the example:$ ./Predictor 0 2 20 2 A
- afterwords you can use the testing-mode by changing the first parameter:
$ ./Predictor 1 2 20 2 A
C++ Class "PTMPredictor" with the following interface (suggestion): void predict(std::vector- useful OpenMS classes
- String, AASequence, FASTAFile
Tasks
Implement prediction of modification state for a given set of peptides and a trained PTM model.
The peptide set can be arbitrary, but should be realistic (e.g. digest a couple of human proteins and save as FASTA file). The trained PTM model can be obtained from the paper (or train yourself [ambitious]).
[22/06/2010] Proposed executable has the following interface:
PTMSimulator -in <my_peptides.FASTA> -out <my_peptides_with_PTMs.FASTA> -mod_p_threshold <float[0-1]> -mod_fraction <float[0-1]> -append
Arguments explained:
'in' - input FASTA file containing peptides/proteins (do not digest)
'out' - modified sequences containing PTMS
'mod_p_threshold' - a threshold for a modification to be accepted
'mod_fraction' - the maximal fraction of the input set which receives a modification. If more sequences receive modification, only use the highest scoring ones
'append' - concatenate the modified set with the original file before writing result to 'out'; be careful not to include unmodified sequences twice
If you have other ideas about how to control the state of modified sequences and/or the total fraction modified, let me know.
In order to ease development, derive your executable from TOPPBase (see OpenMS/source/APPLICATIONS/TOPP/FileConverter.C for an easy example, just put your code into main_() and modify registerOptionsAndFlags_() ).Comments
Copyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors. Ideas, requests, problems regarding Foswiki? Send feedback