SmallWorld Documentation for Future Developer

From DISI
Jump to navigation Jump to search

Introduction

Here is the link to Smallworld's manual

  • Username: ucsf@nextmovesoftware.com
  • Password: <Ask jjiteam@googlegroups.com>

All smallworld databases live in abacus' db3.

Smallworld frontend files are consolidated and located at /export/db3/smallworld_extract/.extract.

Execute command start_smallworld to see options on which smallworld instance to start/restart.

How to Download Smallworld

  1. Ssh to abacus and become root. Prepare directory
     mkdir /opt/nextmove/smallworld-<version> && cd /opt/nextmove/smallworld-<version>
  2. Download Software with this link
    • Username: ucsf@nextmovesoftware.com
    • Password: <Ask jjiteam@googlegroups.com>
  3. Go to releases. Look for smallworld-java-<version>.tar.gz and copy the link address.
  4. Download using wget
     wget --user ucsf@nextmovesoftware.com --password <Ask jjiteam@googlegroups.com> <link address>
  5. Decompress the file
    •  tar -xvf <file_name>

How to Launch Smallworld For The First Time

Prepare Files and Directories

  1. Ssh to abacus and become root
  2. Open a port in the firewall
    firewall-cmd --permanent --add-port=<port_number>/tcp 
    firewall-cmd --reload
  3. Go to smallworld databases directory
    cd /export/db3/
  4. Create a Smallworld directory and go to it
    mkdir <smallworld_dir> && cd <smallworld_dir>
  5. Inside that directory, create these
    • a symbolic link to the smallworld anon map
    ln -s /export/db2/smallworld_anon_21Q4 anon
    • a sub-directory called maps to store databases
    mkdir maps
    • a smallworld config file
    vim smallworld.cfg
    • Add this block of code inside the config file
    [WebApp]
    DepictionUrl=./depict/svg?w=%w&h=%h&smi=%s&qry=%q&cols=%c&cmap=%m
    ResolverUrl=./util/smi2mol?smi=%s
    SearchTimeOut=5000
    SearchAsYouDraw=true
    [Score]
    Name=Atom Alignment
    Description=Aligns and scores each hit relative to the query atom types, differences are categorised as MAJ, MIN, HYB, SUB 
    Class=com.nmsoftware.smallworld.score.SimpleAtomAlignment 
    EnabledByDefault=false
    MappingMode=EditMap
    [Score]
    Name=SMARTS Alignment
    Description=Treat input as SMARTS and align and score each hit relative to the query atom expressions in the input 
    Class=com.nmsoftware.smallworld.score.SmartsAtomAlignment 
    EnabledByDefault=false
    [Score]
    Name=ECFP4
    Description=Extended Connectivity Circular Fingerprint
    Class=com.nmsoftware.smallworld.score.EcfpTanimoto
    Radius=2
    EnabledByDefault=true
    [Score]
    Name=Daylight
    Description=Path based fingerprint
    Class=com.nmsoftware.smallworld.score.DaylightTanimoto
    EnabledByDefault=true
    • SearchTimeOut should be the only option you'll change in the future. It caps search time to lessen strain on machine.
  6. Place any *.anon.maps you've already compiled for testing into the maps sub-directory.

Start Smallworld Instance

  1. Go to smallworld extract directory
    cd /export/db3/smallworld_extract
  2. Create a screen
    screen -S <screen_name>
  3. Export the smallworld directory you've made
    export SWDIR=/export/db3/<smallworld_dir>
  4. Launch java
    java -jar /opt/nextmove/smallworld-<version>/sw-ws-<version>.jar --httpPort=<port_number>

How to Build Smallworld Maps/Databases

Consolidate smiles into /export/db3/temp_mapping.

Use the script located in that directory called smallworld_map_script.

Here is the content of the script:

#!/bin/csh -f
setenv SWDIR /export/db3/public_smallworld_4th_gen/

alias sw 'time java -jar /opt/nextmove/smallworld-5.1/sw.jar'

set target = '*.smi'

foreach i ($target)
	echo '(sw map' $i '-T tmp/ -o '${i}'.anon.map -u '${i}'.anon.unmapped) |& tee '${i}'_log'
	(sw map $i -T tmp/ -o ${i}.anon.map -u ${i}.anon.unmapped) |& tee ${i}_log
	echo 'sw blm -p 0.01 '${i}'.anon.map'
	sw blm -p 0.01 ${i}.anon.map
end

Command Details

setenv sets the environment where the anon master map is

java -jar /opt/nextmove/smallworld-5.1/sw.jar shortened to sw via alias

map will map the smiles based on the current anon master map.

  • -T stores temp files so /tmp does not get overloaded
  • -o output with custom name you like
  • -u will collect the unmapped molecules to send to Nextmove or to use swextend

tee will log the progress of the process

Splitting Databases

[Follow this page if John has asked you to split the database into smaller slices.]

Running SmallWorld from the Command Line

We have been running SmallWorld on n-9-22 as www which uses csh as the default shell. There is a file in /mnt/nfs/soft/www/home/smallworld called run_sw which contains the following commands to set up the SmallWorld Command Line Interface (CLI)

Set the SWDIR environment variable so SmallWorld scripts can access the 12 TB index

   setenv SWDIR /srv/db4

Set the path to the SmallWorld C++ Executables

   set path = ($path /opt/nextmove/bin)

Alias the keyword "sw" to access the Java CLI

   alias sw 'java -jar /opt/nextmove/sw.jar'

Tips for Running SmallWorld on the Command Line

Use sw help to get information about available commands

    sw help