How to run rosetta ab initio modeling

Download

Go to: https://www.rosettacommons.org/software/academic/

Username: Academic_User

Password: Xry3x4

Secondary Structure prediction

First, predict secondary structure (SS) at Robetta.

ab initio

AbinitioRelax.linuxgccrelease -in:file:fasta *.fasta -in:file:frag3 aat000_03_05.200_v1_3 -in:file:frag9 aat000_09_05.200_v1_3 -database /home/yang/software/rosetta/rosetta_database -abinitio:relax -nstruct 1000 -out:pdb -use_filters true -psipred_ss2 *.psipred_ss2 -abinitio::increase_cycles 10 -abinitio::rg_reweight 0.5 -abinitio::rsd_wt_helix 0.5 -abinitio::rsd_wt_loop 0.5 -relax::fast

loop_modeling

loopmodel.linuxgccrelease -in:file:fullatom -in:file:s 3o0g-a_b.pdb -loops:loop_file cdk5.loop_file -loops:frag_sizes 9 3 -loops:frag_files cdk5ex_09_05.200_v1_3 cdk5ex_03_05.200_v1_3 -loops:remodel perturb_ccd -loops:ccd_closure -loops:random_loop -database /home/yang/software/rosetta/rosetta_database/ -mute core.io.database -loops:relax fastrelax -loops:refine refine_kic -loops:extended true -nstruct 1000

Predict Protein-protein Interactions

PPK: docking_prepack_protocol.linuxgccrelease -database /home/yang/software/rosetta/rosetta_database/ -s *.pdb -docking:partners B_A -ignore_unrecognized_res > ppk.log &

LOW: docking_protocol.linuxgccrelease -database /home/yang/software/rosetta/rosetta_database/ -low_res_protocol_only -s 2-22_0001.pdb -docking:partners B_A -spin -dock_pert 3 8 -nstruct 10000 -out:file:scorefile low.sc >low.log &

HIGH: docking_protocol.linuxgccrelease -database /home/yang/software/rosetta/rosetta_database/ -docking_local_refine -s *.pdb -docking:partners B_A -use_input_sc -ex1 -ex2aro -nstruct 10000 &

HIGH_minimization: docking_protocol.linuxgccrelease -database /home/yang/software/rosetta/rosetta_database/ -s *.pdb -docking:partners B_A -ex1 -ex2aro -use_input_sc -no_filters -docking_local_refine -dock_min -mute core.util.prof -out:overwrite -out:file:fullatom -mute core.io.database -nstruct 2000 &

Cluster

cluster.linuxgccrelease -database /opt/rosetta/main/database/ -in:file:s *.pdb -out:file:silent_struct_type binary -cluster:radius -1 -score:weights /opt/rosetta/main/database/scoring/weights/score12.wts -cluster:sort_groups_by_energy -run:shuffle -nstruct 1 -in:file:fullatom > LOG

Simple Score calculation

score: score.linuxgccrelease -database /home/yang/software/rosetta/rosetta_database -in:file:s *.pdb -no_optH -ignore_unrecognized_res > score.log&

scons bin mode=release

Some useful Bash scripts

Run multiple calculations

ROBETTA=../robetta
DATABASE=/opt/rosetta/2019.35/main/database/
AbinitioRelax.linuxgccrelease -in:file:fasta $ROBETTA/*.fasta -in:file:frag3 $ROBETTA/aat000_03_05.200_v1_3 -in:file:frag9 $ROBETTA/aat000_09_05.200_v1_3 -database $DATABASE -abinitio:relax -nstruct 1000 -out:pdb -use_filters true -psipred_ss2 $ROBETTA/*.psipred_ss2 -abinitio::increase_cycles 10 -abinitio::rg_reweight 0.5 -abinitio::rsd_wt_helix 0.5 -abinitio::rsd_wt_loop 0.5 -relax::fast

for ii in $(seq 1 2); do
   mkdir -p run$ii
   cd run$ii
   cp ../template-run-sh run.sh
   ./run.sh >& LOG &
   cd ..
done

Rename multiple run directories

for ((ii=12, jj=14; ii>0, jj>2; ii--, jj--))
do
       echo $ii $jj
done

Sort scores

cat score.fsc | sort -k2 -n | awk '{print $30" "$2}' | sed -n '/description score/q;p' > result.txt

Combine multiple run directories

rm -f tmp
touch tmp
for ii in $(seq 1 22); do
   cat run$ii/result.txt | awk '{print '$ii'" "$1" "$2}' >> tmp
done
sort -k3 -n tmp > result.txt
rm -f tmp

Sort out PDBs from multiple run directories

# NTOP: Number of top-scored (Rosetta score for Ab inito calc) structures you wanna grep
NTOP=5000
mkdir -p top$NTOP
head -$NTOP result.txt | while read line; do
   echo $line
   irun=`echo $line | awk '{print $1}'`
   name=`echo $line | awk '{print $2}'`
   cp run${irun}/${name}.pdb cluster/${irun}.${name}.pdb
done