Pharmit

Copyright (c) David Ryan Koes, University of Pittsburgh and contributors. All rights reserved.

Pharmit is licensed under both the BSD 3-clause license and the GNU Public License version 2. Any use of the code that retains its reliance on the GPL-licensed OpenBabel library is subject to the terms of the GPL2.

Use of the Pharmit code independently of OpenBabel (or any other GPL2 licensed software) may choose between the BSD or GPL licenses.

See the LICENSE file provided with the distribution for more information.

Pharmit is available for searching and creating libraries at http://pharmit.csb.pitt.edu. It generally isn't necessary to build it from source.

BUILDING

CMake is required to build Pharmit. Starting from the src directory: ``` mkdir build cd build

most likely you will have to specify the location of smina nad lemon

cmake .. -DSMINADIR=$HOME/git/smina -DLEMONDIR=/usr/lib/cmake/ make -j12 ```

UBUNTU 22.04

```bash sudo apt install git autoconf automake libtool ghostscript liblemon-dev libeigen3-dev libann-dev bmagic libcgicc-dev libgoogle-perftools-dev libglpk-dev coinor-* libjsoncpp-dev cmake libboost-dev swig libxml2-dev libcairo2-dev libboost-all-dev libcurl4-openssl-dev

git clone https://github.com/FastCGI-Archives/fcgi2.git cd fcgi2 autoreconf -i ./configure sudo make sudo make install cd ..

git clone https://github.com/openbabel/openbabel.git cd openbabel mkdir build cd build sudo cmake .. -DPYTHONBINDINGS=1 -DRUNSWIG=1 -DWITHMAEPARSER=0 -DWITHCOORDGEN=0 sudo make -j12 sudo make install cd ../..

git clone http://git.code.sf.net/p/smina/code smina cd smina mkdir build cd build sudo cmake .. sudo make -j8 sudo cp libsmina.a /usr/local/lib/libsmina.a cd ../..

git clone http://git.code.sf.net/p/pharmit/code pharmit cd pharmit/src mkdir build cd build

change $path/smina the smina directory to reflect the install location from above

sudo cmake .. -DSMINALIB=/usr/local/lib/libsmina.a -DSMINADIR=$path/smina/ -DLEMONDIR=/usr/lib/x86_64-linux-gnu/cmake/lemon/ sudo make -j12 ```

UBUNTU 18.04

``` apt install git autoconf automake libtool ghostscript liblemon-dev libeigen3-dev libann-dev bmagic libcgicc-dev libgoogle-perftools-dev libglpk-dev coinor-* libjsoncpp-dev cmake libboost-dev swig python-dev libxml2-dev libcairo2-dev libboost-all-dev libcurl4-openssl-dev

NOTE: Currently the liblemon-dev package has an incorrect cmake config file that does

not set LEMON_LIBRARY to the correct location. Either edit /usr/lib/cmake/LEMONConfig.cmake

so the correct location (/usr/lib/x8664-linux-gnu/) is set or set LEMONLIBRARY directly

on the cmake commandline

fastcgi

git clone https://github.com/FastCGI-Archives/fcgi2.git cd fcgi2 autoreconf -i ./configure make make install cd ..

openbabel

you can probably use the libopenbabel-dev package, but it doesn't install cmake files by default, so you'd have to provide a custom FindOpenBabel module

In general, I recommend using openbabel 3.0 or later as it fixed a number of bugs.

git clone https://github.com/openbabel/openbabel.git cd openbabel mkdir build cd build

presumably at some point maeparser and coordgen won't be broken...

cmake .. -DPYTHONBINDINGS=1 -DRUNSWIG=1 -DWITHMAEPARSER=0 -DWITHCOORDGEN=0 make -j12 make install cd ../..

smina

git clone http://git.code.sf.net/p/smina/code smina cd smina mkdir build cd build cmake .. make -j8

cd smina/build/linux/release/ make -j8 cp libsmina.a /usr/local/lib/ cd ../../../..

pharmit - ensure SMINA_DIR is the install location above

git clone http://git.code.sf.net/p/pharmit/code pharmit cd pharmit/src mkdir build cd build

/usr/lib/cmake/LEMONConfig.cmake is editted to point LEMONLIBRARY to /usr/lib/x8664-linux-gnu/

cmake .. -DSMINALIB=/usr/local/lib/libsmina.a -DSMINADIR=$HOME/git/smina -DLEMONDIR=/usr/lib/cmake/ make -j12 ```

CENTOS (these instructions may be out of date)

``` yum groupinstall 'Development Tools' yum install git boost-devel autoconf automake libtool make cmake gperftools-libs gperftools-devel ghostscript libcurl-devel

yum install https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm yum install glpk-devel eigen3-devel fcgi coin-or-*

fastcgi

git clone https://github.com/FastCGI-Archives/fcgi2.git cd fcgi2 autoreconf -i ./configure make make install cd ..

jsoncpp

git clone https://github.com/open-source-parsers/jsoncpp.git cd jsoncpp

using the older version mostly to avoid having to upgrade cmake wget

git checkout 1.7.7 mkdir build cd build cmake -DBUILDSHAREDLIBS=1 .. make make install cd ../..

cgicc

wget http://ftp.gnu.org/gnu/cgicc/cgicc-3.2.19.tar.gz tar xvfz cgicc-3.2.19.tar.gz cd cgicc-3.2.19 ./configure make make install cd ..

openbabel

I'm getting the development version since it fixes some serious bugs with aromaticity detection,

but the version in the epel repository is probably fine

git clone https://github.com/openbabel/openbabel.git cd openbabel mkdir build cd build cmake .. make make install cd ../..

eigen

wget http://bitbucket.org/eigen/eigen/get/3.3.5.tar.gz tar xvfz 3.3.5.tar.g cd eigen-eigen-b3f3d4950030/ mkdir build cd build cmake .. make install cd ../..

lemon

wget http://lemon.cs.elte.hu/pub/sources/lemon-1.3.1.tar.gz tar xvfz lemon-1.3.1.tar.gz cd lemon-1.3.1 mkdir build

remove incompatible cmake policy

sed -i '/POLICY/d' CMakeLists.txt cd build cmake .. make make install cd ../..

libann

wget https://www.cs.umd.edu/~mount/ANN/Files/1.1.2/ann1.1.2.tar.gz tar xvfz ann1.1.2.tar.gz cd ann_1.1.2 make linux-g++ cp -r include/ANN /usr/include/ cp lib/libANN.a /usr/lib/libann.a cd ..

bitmagic

git clone https://github.com/tlk00/BitMagic.git cd BitMagic

ubuntu version is actualy 3.7

git checkout v3.8.0

include as part of pharmit - can also install if you wish, but that's not what cmake is expecting

cp -r src/ ~/pharmit/src/bm

smina

git clone http://git.code.sf.net/p/smina/code smina cd smina/build/linux/release/ make cp libsmina.a /usr/local/lib/

pharmit

git clone http://git.code.sf.net/p/pharmit/code pharmit cd pharmit/src mkdir build cmake .. -DCMAKECXXFLAGS=-std=c++0x -DSMINA_DIR=$HOME/smina

lots of warnings about missing directory which can be ignored for now

```

USING

The --help option will provide the following: ``` USAGE: pharmitserver [options] --cmd command [pharma, dbcreate, dbcreateserverdir, dbsearch, server]

OPTIONS: -cmd= - command [pharma, dbcreate, dbcreateserverdir, dbsearch, server] -dbdir= - database directory(s) -dbinfo= - [dbcreateserverdir] JSON file describing database subset -extra-info - Output additional molecular properties. Slower. -file-partition - Partion database slices based on files -help - Display available options (--help-hidden for more) -in= - input file(s) -ligs= - [dbcreateserverdir] Text file listing locations of molecules -logdir= - log directory for server -max-hits= - return at most n results -max-nrot= - maximum allowed rotatable bonds -max-orient= - return at most n orientations of each conformation -max-rmsd= - maximum allowed RMSD; default max allowed by query -max-weight= - maximum allowed molecular weight -min-nrot= - minimum allowed rotatable bonds -min-port= - port for minimization server -min-server= - minimization server address -min-weight= - minimum allowed molecular weight -noindex - [dbcreateserverdir] Do not create indices -nthreads= - utilize n threads; default 1 -out= - output file(s) -pharmaspec= - pharmacophore specification -port= - port for server to listen on -prefixes= - [dbcreateserverdir,server] File of directory prefixes to use for striping. -print - print results -q - quiet; suppress informational messages -receptor= - Receptor file for interaction pharmacophroes -reduceconfs= - return at most n conformations for each molecule -show-query - print query points -singledir= - Specify a single directory to recreate on a dbcreateserverdir command -sort-rmsd - Sort results by RMSD. -timestamp= - Specify timestamp to use for server dirs (for use with single) -unweighted-rmsd - Compute minimal RMSD without radius weights ```

To IDENTIFY the pharmacophore features of a molecule ex.pdb: pharmitserver pharma -in ex.pdb pharmitserver pharma -in ex.pdb -out out.sdf pharmitserver pharma -in ex.pdb -out out.json For interactive, graphical editting of pharmacophore features, try http://pharmit.csb.pitt.edu The saved session file is a json file that can be used as input for Pharmer.

To CREATE a database DB from library.sdf: pharmitserver dbcreate -dbdir DB -in library.sdf If you have multiple disk drives, you can improve performance by striping the database across the drives: pharmitserver dbcreate -dbdir /drive1/DB -dbdir /drive2/DB -in library.sdf If you pre-split the input file things will go faster: pharmitserver dbcreate -dbdir /drive1/DB -dbdir /drive2/DB -file-partition -in library1.sdf -in library2.sdf

To SEARCH a database: pharmitserver dbsearch -dbdir DB -in query.json pharmitserver dbsearch -dbdir DB -in query.sdf pharmitserver dbsearch -dbdir DB -in query.ph4 When setting up a SERVER you create separate directories for each database to search: pharmitserver dbcreateserverdir -ligs molport.ligs -prefixes ../dbprefixes -dbinfo dbinfo.json The ligs files provides the files to add with unique ids and names (see extractsmisubset.py). e.g.: /data22/conformers/541/5416265.sdf.gz 5416265 MolPort-020-216-564 MCULE-1493214818 PubChem-56905170 ZINC000072151660 /data17/conformers/53/534584.sdf.gz 534584 MolPort-000-887-730 MolPort-035-708-781 MCULE-2607511133 MCULE-2857980910 PubChem-19618431 PubChem-90484256 ZINC 000002535004 ZINC02535004 /data08/conformers/18890/188901458.sdf.gz 188901458 MolPort-007-641-322 MCULE-8587443370

The prefixes file specifies the directories (on different drives) to partition the database across. e.g.: /data00/databases /data01/databases /data02/databases

The dbinfo.json file describes the library. e.g.: { "name" : "MolPort", "html" : "MolPort<span><a target=\"_blank\" href=\"https://www.molport.com\" class=\"ui-icon-info ui-icon\"></a></span>", "subdir" : "molport", "prefix" : "MolPort" }

To run a SERVER: pharmitserver server -port 16000 -prefixes /home/dkoes/vendors/dbprefixes -logdir /home/dkoes/log -min-server localhost -min-port 18000

To run a MINIMIZATION server: ~/git/smina/build/linux/release/server -port 18000 -logfile /home/dkoes/log/minlog

To HOST a server, add a fastcgi rule connecting the webserver hosting /web to the storage server running the backend. e.g. (apache): ``` ln -s ~/pharmit/web /var/www/html mkdir /var/www/html/fcgi-bin

apt install apache2 libapache2-mod-php php-mysql apache2-dev

git clone https://github.com/FastCGI-Archives/mod_fastcgi.git cp Makefile.AP2 Makefile

change /usr/local/apache2 to /usr/share/apache2 in Makefile

make make install

create /etc/apache2/mods-available/fastcgi.load

LoadModule fastcgimodule /usr/lib/apache2/modules/modfastcgi.so

fastcgi.conf

AddHandler fastcgi-script .fcgi #FastCgiWrapper /usr/lib/apache2/suexec FastCgiIpcDir /var/lib/apache2/fastcgi

Alias /fcgi-bin/ /var/www/html/fcgi-bin/

SetHandler fastcgi-script Options +ExecCGI

FastCgiExternalServer /var/www/html/fcgi-bin/pharmitserv.fcgi -host 127.0.0.1:16000 -idle-timeout 300 FastCgiExternalServer /var/www/html/fcgi-bin/createlib.fcgi -host 127.0.0.1:11111 -idle-timeout 300

FastCgiConfig -autoUpdate -maxClassProcesses 1

enable

a2enmod fastcgi mkdir /var/www/html/fcgi-bin #this directory has to exist ```

LIBRARY CREATION

The library creation scripts require rdkit to be installed. ``` apt install python3-numpy mysql-server mysql-client python3-mysqldb python3-pip pip3 install flup psutil mysql < ~/pharmit/scripts/pharmit.sql mysql < ~/pharmit/scripts/conformers.sql

also, run the mysql commands:

CREATE USER 'pharmit'@'localhost'; GRANT ALL PRIVILEGES ON pharmit.* TO 'pharmit'@'localhost'; GRANT ALL PRIVILEGES ON conformers.* TO 'pharmit'@'localhost';

INSERT INTO pharmit.users (email,name,maxprivatedbs,maxprivateconfs,maxdbs,maxconfs) VALUES ("guest","Anonymous",0,0,0,10000);

git clone https://github.com/rdkit/rdkit.git cd rdkit mkdir build cd build cmake .. -DPYTHON_EXECUTABLE=/usr/bin/python3 make -j12 ```

BUGS (Not Really)

When working with large datasets spread across multiple hard drives, you will likely run into a number of limits that will result in segmentation faults (type 6 or 11). The following changes need to be in place before building the database, or it will silently create incomplete databases that crash when you search them.

Make sure nofile is set high enough (1000000) in /etc/security/limits.conf

Make sure vm.maxmapcount is set high enough - add the following to /etc/sysctl.conf vm.max_map_count=1000000

Make sure you build against libcurl4-openssl-dev, not gnutls, or you will get fortify_fail errors.

Continuing with curl bugs, DNS timeouts can cause a longjmp error: http://stackoverflow.com/questions/9191668/error-longjmp-causes-uninitialized-stack-frame to resolve these rebuild curl with --enable-ares

You can ignore all the curl issues entirely if you build with -DSKIP_REGISTERZINC which is probably what you want unless you are using the full ZINC database.

Some versions of eigen3 will trigger an assert, OBJECTALLOCATEDONSTACKISTOOBIG, at compile time. This can be ignored in the latest versions of eigen3 with -DEIGENDISABLESTACKSIZEASSERT If your version hasn't had this patch applied (which is currently true with Ubuntu 14.04), then you will have to modify Eigen/src/Core/DenseStorage.h appropriately.

Currently, on Ubuntu 14.04, g++ 4.8 experiences an internal compiler error. Use g++-4.6 to get around this.

BUGS (Really)

Send any bug reports to dkoes@pitt.edu with a complete test case.

CITING

Please use the citation from http://pubs.acs.org/doi/abs/10.1021/ci200097m

Default Pharmacophore Definitions

```` Aromatic 18 0 1.1 1 0.1 a1aaaaa1 a1aaaa1

HydrogenDonor 1 1 0.5 1 0.1 [#7!H0&!$(N-SX4(=O)CX4(F)F)] [#8!H0&!$([OH][C,S,P]=O)] [#16!H0]

HydrogenAcceptor 89 2 0.5 1 0.1 [#7&!$([nX3])&!$([NX3]-=[!#6])&!$([NX3]-[a])&!$([NX4])&!$(N=C([C,N])N)] [$([O])&!$(OX2C=O)&!$((~a)~a)]

PositiveIon 7 3 0.75 0 0.1 [+,+2,+3,+4] $(CC)N [$(C(N)(N)=N)] [$(n1cc[nH]c1)]

NegativeIon 8 4 0.75 0 0.1 [-,-2,-3,-4] C(=O)[O-,OH,OX1] [$(S,P[O-,OH,OX1])] c1[nH1]nnn1 c1nn[nH1]n1 C(=O)N[OH1,O-,OX1] C(=O)N[OH1,O-] CO(=N[OH1,O-]) [$(N-SX4(=O)CX4(F)F)]

Hydrophobic 6 5 1 0 2 a1aaaaa1 a1aaaa1 [$([CH3X4,CH2X3,CH1X2,F,Cl,Br,I])&!$(*[CH3X4,CH2X3,CH1X2,F,Cl,Br,I])] [$(([CH3X4,CH2X3,CH1X2,F,Cl,Br,I])[CH3X4,CH2X3,CH1X2,F,Cl,Br,I])&!$(*([CH3X4,CH2X3,CH1X2,F,Cl,Br,I])([CH3X4,CH2X3,CH1X2,F,Cl,Br,I])[CH3X4,CH2X3,CH1X2,F,Cl,Br,I])]([CH3X4,CH2X3,CH1X2,F,Cl,Br,I])[CH3X4,CH2X3,CH1X2,F,Cl,Br,I] *([CH3X4,CH2X3,CH1X2,F,Cl,Br,I])([CH3X4,CH2X3,CH1X2,F,Cl,Br,I])[CH3X4,CH2X3,CH1X2,F,Cl,Br,I] [C&r3]1~[C&r3]~[C&r3]1 [C&r4]1~[C&r4]~[C&r4]~[C&r4]1 [C&r5]1~[C&r5]~[C&r5]~[C&r5]~[C&r5]1 [C&r6]1~[C&r6]~[C&r6]~[C&r6]~[C&r6]~[C&r6]1 [C&r7]1~[C&r7]~[C&r7]~[C&r7]~[C&r7]~[C&r7]~[C&r7]1 [C&r8]1~[C&r8]~[C&r8]~[C&r8]~[C&r8]~[C&r8]~[C&r8]~[C&r8]1 [CH2X4,CH1X3,CH0X2]~[CH3X4,CH2X3,CH1X2,F,Cl,Br,I] [$([CH2X4,CH1X3,CH0X2]~[$([!#1]);!$([CH2X4,CH1X3,CH0X2])])]~[CH2X4,CH1X3,CH0X2]~[CH2X4,CH1X3,CH0X2] [$([CH2X4,CH1X3,CH0X2]~[CH2X4,CH1X3,CH0X2]~[$([CH2X4,CH1X3,CH0X2]~[$([!#1]);!$([CH2X4,CH1X3,CH0X2])])])]~[CH2X4,CH1X3,CH0X2]~[CH2X4,CH1X3,CH0X2]~[CH2X4,CH1X3,CH0X2] [$([S]~[#6])&!$(S~[!#6])] ````