How to setup CS-HM-Rosetta calculations

Submitted by snerli on Mon, 02/19/2018 - 08:00

CS-HM-Rosetta also known as Chemical Shift-Homology Modeling-Rosetta is a protocol that makes use of structural restraints derived from homologous proteins together with any additional NMR data within CS/RARSEC/AutoNOE-Rosetta protocols to model high-resolution structures [1].

Source availability:

1. Rosetta must be installed first. It is available to academic and non-academic users under the license (see here).
2. The scripts (CS-HM-Rosetta) required to obtain structural restraints from homologous proteins are available under the downloads folder within toolbox version 3.5.

Steps to run CS-HM-Rosetta:

Step1: First, an alignment file (in .hhr format) needs to be generated using HHpred. To generate the alignment, a webserver (see here) or a standalone HHsearch (available here) can be used. If a standalone software is used, the following command needs to be executed from within hhsearch directory for alignment generation. The -i option is used to provide fasta file, -d option is used to provide a database file (this file is obtained by concatenation of all the hhm files from pdb70 database available here).

$ ./hhsearch -i aLP.fasta -d hhsearch_db.hhm -o aLP.hhr

Step2: After the generation of an alignment file (.hhr), the rosetta_cm.conf present within the CS-HM-Rosetta/cm_scripts directory must be updated to reflect the compiler requirements. Below is the example of the .conf file

boinc_tag           rosetta_cm
method              loopbuild_threading loopbuild_threading_cst_relax

# alignment filtering options
max_templates       20
max_pct_id          1.0
max_template_pct_id 1.0
max_e_value         1000.0
mini_compile_mode   release

# If MACOS, then use the clang compiler
mini_compiler       clang
# If LINUX, then use the gcc compiler
#mini_compiler       gcc

# standard boinc priority
priority 0

Here, max_templates represent the maximum number of templates to include, max_pct_id allows defining maximum percentage identity, max_template_pct_id represents maximum percentage identity of templates to query sequence, max_e_value is the threshold to filter based on HHsearch e-values, mini_compile_mode is the compilation mode (either release or debug), mini_compiler tells the protocol to use either gcc compiler (for UNIX/LINUX systems) or clang compiler (for MACOS).

Step3: Once the rosetta_cm.conf is adjusted, the following commands need to be executed to obtain the distance restraints:

$ cd CS-HM-Rosetta/cm_scripts/bin
$ ./predict_distances.pl ../../input/rtt.hhr ../../input/rtt.fasta -aln_format hhsearch -outfile rtt.cst

During the execution of predict distances script, a series of pdb structure are downloaded for estimating the distance restraints.

Step4: The distance restraints obtained in Step3 are applied during structure calculation performed using either of the protocols listed below:

Tutorial: CS-Rosetta
Tutorial: RASREC
Tutorial: AutoNOE

Example input and expected output files are provided under the CS-HM-Rosetta directory within the toolbox.

References

  1. [Thompson2012] Thompson J, Sgourakis NG, Liu G, Rossi P, Tang Y, Mills JL, Szyperski T, Montelione GT, Baker D. 
    2012.  Accurate protein structure modeling using sparse NMR data and homologous structure information. Proceedings of the National Academy of Sciences. 109(25)