Before starting make sure you have the following packages installed:
- Rosetta (from RosettaCommons)
- rosettascripts (from this repo)
The riboswitch that we are modeling here is derived from the yybP-ykoY aptamer of Xanthomoinas oryzae (Suddala, Nat. Commun., 2019).
The RNA sequence and secondary structure annotation are provided as inputs for de novo modeling:
mn_riboswitch.fasta
>mn_riboswitch A:1-57 B:1-48
auccuuggggaguagccugcuuucuucggaaagcgccuguaucaacauacucggcua,uagccguggugcaggcaacggcgaaagccgucuggcgagaccagggau
mn_riboswitch_secstruct.txt
((((((((......(((.((((((....))))))((((((((((.......((((((,)))))))))))))))).(((((....)))))..)))....))))))))
The FARAFAR2 protocol of Rosetta includes improved base-pair sampling, and an updated fragment library and socring function (Watkins et al., Structure 2020). Models are calculated with rna_denovo
. To run rna_denovo
on multiple cores in parallel rosettascripts provides a utility script submitJobs
to kickstart the simulations. First write the call to rna_denovo
into a file.
FARFAR2.txt
rna_denovo.linuxgccrelease
-nstruct 1000
-fasta mn_riboswitch.fasta
-secstruct_file mn_riboswitch_secstruct.txt
-silent mn_riboswitch.out
-minimize_rna true
-cycles 20000
The flags used here are:
-nstruct
number of models to compute on each core-fasta
path to fasta sequence file-secstruct_file
path to secondary structure in dot-bracket notation-silent
output silentfile where models are stored-minimize_rna
whether to refinement in the output structures with the high-resolution Rosetta potential-cycles
number of Monte Carlo cycles)
Then, start the simulations on as many cores as you like (here: 12) and write the models to a silentfile in the specified output directory (here: "denovo")
submitJobs -i FARFAR2.txt -d denovo -p 12
At any time during the modeling you can check the number of generated models by running extract_pdb
in dry mode (-e false
prevents extraction of any PDB file):
extract_pdb -d denovo -e false
You can extract the top scoring PDB models (here: n=10) from the Rosetta generated silentfiles (set the -m
flag to true if you want to merge them into a single PDB file)
extract_pdb -d denovo -n 10 -m true