Large-scale TC Calculations

Written by Jiankun Lyu, 20170918

Please run Large-scale SMILES Requesting and Fingerprints Converting first to get the SMILES and fingerprints files, then run the TC calculations below.

1) Run locally

/mnt/nfs/home/jklyu/zzz.github/ChemInfTools/utils/cal_Tc_matrix_uint16/cal_Tc_matrix_uint16 (1)first_fingerprint_file (2)first_smiles_file (3)first_count_file (4)second_fingerprint_file (5)second_smiles_file (6)second_count_file

Here is one example for calculating the best TC value to all aminergic knowns for each top docking cluster head:

nohup /mnt/nfs/home/jklyu/zzz.github/ChemInfTools/utils/cal_Tc_matrix_uint16/cal_Tc_matrix_uint16 /path/to/the/cluster_head_uint16.fp /path/to/the/cluster_head.smi /path/to/the/cluster_head_uint16.count /mnt/nfs/reshwork/jklyu/fp_db/all_aminergic/all_aminergic_uint16.fp /mnt/nfs/reshwork/jklyu/fp_db/all_aminergic/all_aminergic.smi /mnt/nfs/reshwork/jklyu/fp_db/all_aminergic/all_aminergic_uint16.count > log &

2) Submit to the queue(calculate Tc to Agent)

2.1) split the smiles and fingerprints files

python ~jklyu/work/cluster/script/make_chunks_for_file_new.py cluster_head.smi cluster_head.smi 500 .
python ~jklyu/work/cluster/script/make_chunks_for_file_new.py cluster_head.fp cluster_head.fp 500 .

2.2) copy precalculated fingerprints to each machine

mkdir cp_fp_to_nodes
cd cp_fp_to_nodes/
csh ~jklyu/work/cluster/script/cp_files_2_scratch.csh
cd ../

2.3) submit jobs to the queue

csh ~jklyu/work/cluster/script/run.Tc.matrix.new.csh ~jklyu/work/cluster/script

2.4) collect data

python ~jklyu/work/cluster/script/combine_and_find_closest_neighbor.py 502 cluster_head_