c 2017

Size: px

Start display at page:

Download "c 2017"

まいかこうだ
5 years ago
Views:

1 CzeekS ver. 1.2

2 c 2017

3 ver. 1.2 Dragon7 shrink learn status

4 1 1 2 CzeekS OpenBabel CGBVS Tanimoto CGBVS DB cgbvs 18

5 1 - ChEMBL CGBVS Chemical Genomics-Based Virtual ScreeningCzeekS CGBVS CGBVS MACCS 2 CzeekS CzeekS CzeekS OpenMP CPU 1 CzeekS CPU 4 CPU (Intel, AMD) 16Gb HDD 10Gb OS CentOS5.x or 6.x 64bit (Linux 2.6) DRAGON6.0.38, Dragon7.0x OpenBabel CPU Intel Xeon E Gb 20h 10m Intel Core i Gb 66h 52m AMD Phenom X6 1055T 6 8Gb 70h 40m 1

6 1 CPU Ion Nuclear GPCR Kinase Channel Receptor Intel Xeon E GB Intel Core i GB Protease Transporter 10m 14s 20m 29s 7m 23s 4m 38s 10m 2s 3m 23s 9m 30s 18m 57s 7m 41s 5m 26s 9m 54s 3m 6s 2

7 2 CzeekS 2.1 CzeekS ******.tgz tar /usr/local czeeks /home/czeeks /home/czeeks $ tar xvzf CzeekS_ ******. tgz CGBVS / CGBVS / exec / CGBVS / exec / protein. lst CGBVS / exec /2 D_7_910_smi. drs CGBVS / exec / cgbvs CGBVS / exec / calc_dragon.sh CGBVS / exec /2 D_7_910_sdf. drs CGBVS / exec / SVMlearn CGBVS / exec / minfo czeeks license.dat /home/czeeks/cgbvs/exec CGBVS/ example H3 mols.csv H3 mols.fp H3 mols.sdf H3 mols.smi H3 positive.csv gpcr.csv positive.csv sample mols.csv sample mols.fp sample mols.sdf sample mols.smi training mols.csv training mols.fp training mols.sdf training mols.smi exec 2D sdf.drs 2D smi.drs 2D 894 sdf.drt 2D 894 smi.drt SVMlearn calc FP MACCS calc dragon.sh calc dragon7.sh cgbvs czeeks license.dat minfo protein.lst gpcr sample.db H3 H3 H3 SD H3 SMILES H3 GPCR SDF SMILES SD SMILES DRAGON7 DRAGON7 DRAGON6 DRAGON6 SVM MACCS DRAGON6 DRAGON7 CGBVS DB 3

8 CGBVS/exec minfo SHA1 2. $ cd CGBVS / exec $./ minfo a5866b20b7b4a1da0ac4406dcf7f40b963903c34 // 3. 2 a CGBVS czeeks license.dat b CGBVS LICENSE par bashrc $ export CGBVS =/ home / czeeks / CGBVS / exec $ export PATH = $PATH : $CGBVS $ export LD_LIBRARY_PATH =/ usr / local / lib : $LD_LIBRARY_PATH $ export DRAGON6 =/ usr / local / bin // DRAGON6 $ export DRAGON7 =/ usr / local / bin // DRAGON7 DRAGON6/DRAGON7 DRAGON dragon6shell dragon7shell czeeks license.dat ${CGBVS} CGBVS LICENSE 4

9 2.4 OpenBabel CzeekS MACCS calc FP MACCS SD SMILES OpenBabel 1. cmake OpenBabel cmake cmake yum install cmake 2. OpenBabel OpenBabel GPL v2 URL Open Babel tar openbabel openbabel $ mkdir build // $ cd build $ cmake../ // c m a k e $ make // OpenBabel $ su // # make install // CzeekS OpenBabel OpenBabel 5

10 3 3.1 CGBVS CzeekS CzeekS.db DB ChEMBL GPCR 4 CGBVS SVM SVM CzeekS CGBVS 2 1 SVM CzeekS CGBVS cgbvs status DB DB $ cgbvs status gpcr_ sample. db [ compound ] Dragon6 v // # of data = // # of descriptors = 894 // [ protein ] PROFEAT 2011 // # of data = 859 // # of descriptors = 1080 // [ fingerprint ] 6

11 MACCS // # of data = // [ interactions ] # of positive interactions = // # of negative interactions = 0 // [ details of models ] # of sampled positive interactions = // id nsv dim C gamma 5- fold CV id id nsv C gamma SVM accuracy cgbvs status -p $ cgbvs status - p gpcr_ sample. db protein ID positive negative accession name 5 HT1A_ HUMAN P hydroxytryptamine receptor 1 A 5 HT1B_ HUMAN P hydroxytryptamine receptor 1 B 5 HT1D_ HUMAN P hydroxytryptamine receptor 1 D 5 HT1E_ HUMAN 74 0 P hydroxytryptamine receptor 1 E 5 HT1F_ HUMAN P hydroxytryptamine receptor 1 F 5 HT2A_ HUMAN P hydroxytryptamine receptor 2 A 5 HT2B_ HUMAN P hydroxytryptamine receptor 2 B 5 HT2C_ HUMAN P hydroxytryptamine receptor 2 C 5 HT4R_ HUMAN Q hydroxytryptamine receptor 4 5 HT5A_ HUMAN P hydroxytryptamine receptor 5 A 5 HT6R_ HUMAN P hydroxytryptamine receptor 6 5 HT7R_ HUMAN P hydroxytryptamine receptor 7 A4_ HUMAN P05067 Amyloid beta A4 protein AA1R_ HUMAN P30542 Adenosine receptor A1 AA2AR_ HUMAN P29274 Adenosine receptor A2a AA2BR_ HUMAN P29275 Adenosine receptor A2b AA3R_ HUMAN P33765 Adenosine receptor A3 protein ID ID ID accession UniProt( ID positive DB negative 7

12 3.2 SD DB CzeekS DRAGON6 exec SMILES DRAGON6 SMILES OpenBabel SD SMILES $ babel - isdf sample_ mols. sdf - osmi sample_ mols. smi // S M I L E S $ calc_ dragon. sh sample_ mols. smi > output. csv $ cat output. csv ZINC , ,8.522,24.952,38.109,25.091, ZINC , ,8.416,21.796,32.563,22.216, ZINC , ,7.152,25.928,42.138,27.228, ZINC , ,10.941,21.362,32.153,21.784, ZINC , ,6.778,22.928,39.138,24.228, // C S V ID, 1, 2, 1 1 ID calc dragon.sh cgbvs predict CzeekS sample mols.csv β2 $ cgbvs predict gpcr_ sample. db ADRB2_ HUMAN sample_ mols. csv compound ADRB2_ HUMAN ZINC ZINC ZINC ZINC

13 ZINC ZINC ZINC CGBVS DB 3 ID 4 3 ID cgbvs status p 3 ID ID 1 2 $ cgbvs predict gpcr_ sample. db ADRB1_ HUMAN, ADRB2_ HUMAN sample_ mols. csv compound ADRB1_ HUMAN ADRB2_ HUMAN ZINC ZINC ZINC ZINC ZINC ZINC ZINC % $ cgbvs predict gpcr_ sample. db ADA %, ADR % sample_ mols. csv compound ADA1A_ HUMAN ADA1B_ HUMAN ADA1D_ HUMAN ADA2A_ HUMAN ADA2B_ HUMAN ADA2C_ HUMAN ADRB1_ HUMAN ADRB2_ HUMAN ADRB3_ HUMAN ZINC ZINC ZINC cgbvs predict CGBVS -d 9

14 SVM $ cgbvs predict - d gpcr_ sample. db ADR % sample_ mols. csv compound ADRB1_ HUMAN ADRB2_ HUMAN ADRB3_ HUMAN ZINC ZINC ZINC ZINC ZINC ZINC ZINC v $ cgbvs predict - v gpcr_ sample. db ADR % sample_ mols. csv compound protein probability score ZINC ADRB1_ HUMAN ZINC ADRB2_ HUMAN ZINC ADRB3_ HUMAN ZINC ADRB1_ HUMAN ZINC ADRB2_ HUMAN ZINC ADRB3_ HUMAN ZINC ADRB1_ HUMAN ZINC ADRB2_ HUMAN ZINC ADRB3_ HUMAN CGBVS CGBVS cgbvs predict all DB 1 -a cgbvs status -p sample mols.csv ZINC ID $ grep ZINC sample_ mols. csv > test. csv $ cgbvs predict - v gpcr_ sample. db all test. csv 10

15 compound protein probability score ZINC HT1A_ HUMAN ZINC HT1B_ HUMAN ZINC HT1D_ HUMAN ZINC HT1E_ HUMAN ZINC HT1F_ HUMAN ZINC HT2A_ HUMAN ZINC HT2B_ HUMAN ZINC HT2C_ HUMAN ZINC HT4R_ HUMAN ZINC HT5A_ HUMAN ZINC HT6R_ HUMAN ZINC HT7R_ HUMAN ZINC A4_ HUMAN ZINC AA1R_ HUMAN v ID $ cgbvs predict - v gpcr_ sample. db all test. csv > out $ sort - k3 - nr out head ZINC MTR1A_ HUMAN ZINC TSHR_ HUMAN ZINC GRM2_ HUMAN ZINC MTR1B_ HUMAN ZINC HRH3_ HUMAN ZINC HT1E_ HUMAN ZINC CCR6_ HUMAN ZINC NPY2R_ HUMAN ZINC GRM5_ HUMAN ZINC ARBK1_ HUMAN ID MTR1A HUMAN MTR1B HUMAN $ cgbvs status - p gpcr_ sample. db grep - e " MTR1..*" MTR1A_ HUMAN P48039 Melatonin receptor type 1 A MTR1B_ HUMAN P49286 Melatonin receptor type 1 B 3.4 Tanimoto CzeekS Tanimoto Similarity Tanimoto DB Tanimoto cgbvs predict -s 11

16 $ calc_fp_maccs sample_mols. sdf test.fp // test. fp sample mols. f p $ cgbvs predict - s gpcr_ sample. db ADRB2_ HUMAN test. fp compound ADRB2_ HUMAN ZINC ZINC ZINC ZINC ZINC ZINC ZINC ZINC ZINC test.fp $ head sample_ mols. fp ZINC , ZINC , ZINC , ZINC , ZINC , ID 2 1 n 1 n 12

17 4 CGBVS 4.1 CGBVS CSV training mols.csv $ head training_ mols. csv 250, ,8.778,30.037,43.407,31.387,47.870,0.699,1.009,0.730, , ,7.087,55.690,89.934,58.768, ,0.619,0.999,0.653, , ,6.837,33.058,51.032,35.597,57.690,0.636,0.981,0.685, , ,7.439,37.354,55.243,39.637,62.145,0.667,0.986,0.708, , ,6.537,31.858,54.767,34.074,62.893,0.579,0.996,0.620, 7139, ,8.019,33.853,49.691,35.938,55.302,0.677,0.994,0.719, , ,8.249,19.482,31.585,20.352,35.349,0.628,1.019,0.657, , ,8.761,43.782,64.774,45.995,71.368,0.684,1.012,0.719, , ,8.313,27.316,41.243,28.665,46.053,0.666,1.006,0.699, , ,6.537,31.858,54.767,34.074,62.893,0.579,0.996,0.620, 3 1 ID 2 training mols.smi DRAGON6 gpcr.csv $ head gpcr. csv 5 HT1A_HUMAN, , , ,3.5545, , 5 HT1B_HUMAN, , , , , , 5 HT1D_HUMAN, , , , , , 5 HT1E_HUMAN, , , , , , 5 HT1F_HUMAN, , , , , , 5 HT2A_HUMAN, , , , , , 5 HT2B_HUMAN, ,1.6632,2.9106,4.3659, , 5 HT2C_HUMAN,5.8952, , , , , 5 HT4R_HUMAN, , , , ,5.6701, 5 HT5A_HUMAN, , , , , , 13

18 PROFEAT FASTA PRO- FEAT CzeekS ID UniProt ID * HUMAN UniProtID positive.csv $ head positive. csv , NPBW1_ HUMAN , ARBK1_ HUMAN , CRFR1_ HUMAN , FAK2_ HUMAN , CCR6_ HUMAN , NTR1_ HUMAN , FAK2_ HUMAN , OX1R_ HUMAN , PTAFR_ HUMAN , ADRB2_ HUMAN 1 ID 2 ID - ChEMBL - 30µM 4.2 DB 3 CGBVS DB training mols.csv, gpcr.csv, positive.csv $ cgbvs create training.db // D B $ cgbvs import training.db training_mols. csv compound // import training_ mols. csv $ cgbvs import training.db gpcr. csv protein // import gpcr. csv $ cgbvs import training.db positive. csv positive // import positive. csv DB DB gbvs create DB SVM CGBVS

19 3.4 CzeekS DB Tanimoto $ cgbvs import training. db training_ mols. fp fingerprint import training_ mols. fp MACCS DB CGBVS gbvs status -p -a cgbvs add H3 100 H3 mols.sdf H3 mols.csv H3 positive.csv $ cgbvs add training.db H3_mols. csv compound // import H3_ mols. csv $ cgbvs add training.db H3_positive. csv positive // import H3_ positive. csv 4.4 DB SVM cgbvs learn $ cgbvs learn -C 99% -P 99% -n -f training.db 5 output input_ 1 output input_ 2 output input_ 3 output input_ 4 output input_ 5 $ SVMlearn - c 3 - g input_ 1 model_ 1 15

20 itr nsv vkkt Objective E E E c -g SVM -c SVM C CzeekS SVM RBF(Radial Basis Function) -g RBF C=3 γ=0.003 SVM C f SVM $ cgbvs learn - f training. db 5 output input_ 1 output input_ 2 output input_ 3 output input_ 4 output input_ 5 SVM $ SVMlearn -c 3 -g input_1 model_1 // 1 $ SVMlearn -c 3 -g input_2 model_2 // 2 $ SVMlearn -c 3 -g input_3 model_3 // 3 $ SVMlearn -c 3 -g input_4 model_4 // 4 $ SVMlearn -c 3 -g input_5 model_5 // 5 model 1 model 5 5 DB $ cgbvs add_model training.db model_1 1 // model 1 id=1 $ cgbvs add_model training.db model_2 2 // model 2 id=2 $ cgbvs add_model training.db model_3 3 // model 3 id=3 $ cgbvs add_model training.db model_4 4 // model 4 id=4 16

21 $ cgbvs add_model training.db model_5 5 // model 5 id=5 cgbvs status SVM input 1 #!/ bin /sh for c in ; do for g in ; do echo -ne $c "\t"$g "\t" SVMlearn - c $c - g $g input_ 1 model_ 1 grep cross - validation awk { print $6 } done done SVM C=1, 3, 10, 30, γ=0.001, 0.003, 0.01, 0.03, C C model 1 model 5 DB 17

22 5 cgbvs cgbvs < > [< >] < > add, add model, comment, create, delete, del model, import, learn, predict, status, shrink add : cgbvs add <db > < > < > CSV db < > compound protein positive negative fingerprint add model : SVM cgbvs add model <db > < > <ID > SVM ID db ID ID SVMlearn -l libsvm svm-train 18

23 -l :libsvm comment : cgbvs comment <db > < > < > <db > db < > compound protein positive negative fingerprint create : db cgbvs create [ ] <db > db db import -c <arg> <arg> -p <arg> <arg> -i <arg> <arg> -n <arg> <arg> -f <arg> <arg> <arg> CSV delete : cgbvs delete <db > < > 19

24 <db > db < > compound protein positive negative fingerprint del model : SVM db cgbvs del model <db > < ID> <db > db < ID> SVM cgbvs status < ID> all SVM import : db cgbvs import <db > < > < > CSV db < > compound protein positive negative fingerprint add db < > db import -m <arg> <arg> 20

25 learn : cgbvs learn [ ] <db > < > db SVM db SVM SVM < > SVM db -c <arg> SVM C 10 -g <arg> RBF v <arg> 5 -s <arg> 1 -C <arg> -P <arg> 2 <arg> <arg> % -n -r 2 SVM -f SVMlearn -l LIBSVM predict : cgbvs predict [ ] <db > < ID> < > <db > CGBVS < ID> < > 21

26 < ID> % all db ID status -p -a -s Tanimoto -d SVM -v -n <arg> <arg> ID status : db cgbvs status [ ] <db > db -c ID -p ID -a ID -p 1 -a predict ID ID shrink : db cgbvs shrink <db > db db 22

目次 1. はじめに CzeekS のインストールと設定アーカイブの展開とライセンスファイルの配置環境変数の設定 OpenBabel の設定化合物スクリーニングとターゲット予測 CGBVS

目次 1. はじめに CzeekS のインストールと設定アーカイブの展開とライセンスファイルの配置環境変数の設定 OpenBabel の設定化合物スクリーニングとターゲット予測 CGBVS 株式会社京都コンステラテクノロジーズ CzeekS マニュアル 2013/2/26 目次 1. はじめに... 1 2.CzeekS のインストールと設定... 2 2-1. アーカイブの展開とライセンスファイルの配置... 2 2-2. 環境変数の設定... 2 2-3.OpenBabel の設定... 3 3. 化合物スクリーニングとターゲット予測... 4 3-1.CGBVS モデル...