Porting this vignette to knitr is causing some formatting issues. This will be resolved soon.

Email contact: thomas.girke@ucr.edu


ChemmineR is a cheminformatics package for analyzing drug-like small molecule data in R. Its latest version contains functions for efficient processing of large numbers of small molecules, physicochemical/structural property predictions, structural similarity searching, classification and clustering of compound libraries with a wide spectrum of algorithms.

ChemmineR environment with its add-on packages and selected functionalities

ChemmineR environment with its add-on packages and selected functionalities

In addition, ChemmineR offers visualization functions for compound clustering results and chemical structures. The integration of chemoinformatic tools with the R programming environment has many advantages, such as easy access to a wide spectrum of statistical methods, machine learning algorithms and graphic utilities. The first version of this package was published in Cao, Charisi, Cheng, Jiang, and Girke (2008) . Since then many additional utilities and add-on packages have been added to the environment and many more are under development for future releases (Figure 2; (Backman, Cao, and Girke, 2011; Wang, Backman, Horan, and Girke, 2013)).

Getting Started


The R software for running ChemmineR can be downloaded from CRAN (http://cran.at.r-project.org/). The ChemmineR package can be installed from R using the bioLite install command.

 source("http://bioconductor.org/biocLite.R") # Sources the biocLite.R installation script. 
 biocLite("ChemmineR") # Installs the package. 

Loading the Package and Documentation

 library("ChemmineR") # Loads the package
 library(help="ChemmineR") # Lists all functions and classes 
 vignette("ChemmineR") # Opens this PDF manual from R 

Five Minute Tutorial

The following code gives an overview of the most important functionalities provided by ChemmineR. Copy and paste of the commands into the R console will demonstrate their utilities.

Create Instances of SDFset class:

 sdfset <- sdfsample
 sdfset # Returns summary of SDFset 
## An instance of "SDFset" with 100 molecules
 sdfset[1:4] # Subsetting of object
## An instance of "SDFset" with 4 molecules
 sdfset[[1]] # Returns summarized content of one SDF
## An instance of "SDF"
## <<header>>
##                             Molecule_Name 
##                                  "650001" 
##                                    Source 
##                  "  -OEChem-07071010512D" 
##                                   Comment 
##                                        "" 
##                               Counts_Line 
## " 61 64  0     0  0  0  0  0  0999 V2000" 
## <<atomblock>>
##           C1      C2  C3  C5  C6  C7  C8  C9 C10 C11 C12 C13 C14 C15 C16
## O_1   7.0468  0.0839   0   0   0   0   0   0   0   0   0   0   0   0   0
## O_2  12.2708  1.0492   0   0   0   0   0   0   0   0   0   0   0   0   0
## ...      ...     ... ... ... ... ... ... ... ... ... ... ... ... ... ...
## H_60  1.8411 -1.5985   0   0   0   0   0   0   0   0   0   0   0   0   0
## H_61  2.6597 -1.2843   0   0   0   0   0   0   0   0   0   0   0   0   0
## <<bondblock>>
##      C1  C2  C3  C4  C5  C6  C7
## 1     1  16   2   0   0   0   0
## 2     2  23   1   0   0   0   0
## ... ... ... ... ... ... ... ...
## 63   33  60   1   0   0   0   0
## 64   33  61   1   0   0   0   0
## <<datablock>> (33 data items)
##                       "650001"                            "1" 
##                          "700"                            "7" 
##                          "..."
 view(sdfset[1:4]) # Returns summarized content of many SDFs, not printed here 
 as(sdfset[1:4], "list") # Returns complete content of many SDFs, not printed here 

An SDFset is created during the import of an SD file:

 sdfset <- read.SDFset("http://faculty.ucr.edu/ tgirke/Documents/R_BioCond/Samples/sdfsample.sdf") 

Miscellaneous accessor methods for SDFset container:

 header(sdfset[1:4]) # Not printed here
##                             Molecule_Name 
##                                  "650001" 
##                                    Source 
##                  "  -OEChem-07071010512D" 
##                                   Comment 
##                                        "" 
##                               Counts_Line 
## " 61 64  0     0  0  0  0  0  0999 V2000"
 atomblock(sdfset[1:4]) # Not printed here 
##         C1     C2 C3 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16
## O_1  7.047 0.0839  0  0  0  0  0  0   0   0   0   0   0   0   0
## O_2 12.271 1.0492  0  0  0  0  0  0   0   0   0   0   0   0   0
## O_3 12.271 3.1186  0  0  0  0  0  0   0   0   0   0   0   0   0
## O_4  7.913 2.5839  0  0  0  0  0  0   0   0   0   0   0   0   0
bondblock(sdfset[1:4]) # Not printed here 
##   C1 C2 C3 C4 C5 C6 C7
## 1  1 16  2  0  0  0  0
## 2  2 23  1  0  0  0  0
## 3  2 27  1  0  0  0  0
## 4  3 25  1  0  0  0  0
 datablock(sdfset[1:4]) # Not printed here 
##                       "650001"                            "1" 
##                          "700"                            "7"

Assigning compound IDs and keeping them unique:

 cid(sdfset)[1:4] # Returns IDs from SDFset object
## [1] "CMP1" "CMP2" "CMP3" "CMP4"
 sdfid(sdfset)[1:4] # Returns IDs from SD file header block
## [1] "650001" "650002" "650003" "650004"
 unique_ids <- makeUnique(sdfid(sdfset))
## [1] "No duplicates detected!"
 cid(sdfset) <- unique_ids 

Converting the data blocks in an SDFset to a matrix:

 blockmatrix <- datablock2ma(datablocklist=datablock(sdfset)) # Converts data block to matrix 
 numchar <- splitNumChar(blockmatrix=blockmatrix) # Splits to numeric and character matrix 
 numchar[[1]][1:2,1:2] # Slice of numeric matrix 
## 650001               650001                              1
## 650002               650002                              1
 numchar[[2]][1:2,10:11] # Slice of character matrix 
## 650001 "C23H28N4O6"             
## 650002 "C18H23N5O3"             
##        PUBCHEM_OPENEYE_CAN_SMILES                                     
## 650001 "CC1=CC(=NO1)NC(=O)CCC(=O)N(CC(=O)NC2CCCC2)C3=CC4=C(C=C3)OCCO4"
## 650002 "CN1C2=C(C(=O)NC1=O)N(C(=N2)NCCCO)CCCC3=CC=CC=C3"

Compute atom frequency matrix, molecular weight and formula:

 propma <- data.frame(MF=MF(sdfset), MW=MW(sdfset), atomcountMA(sdfset))
 propma[1:4, ] 
##                 MF    MW  C  H N O S F Cl
## 650001  C23H28N4O6 456.5 23 28 4 6 0 0  0
## 650002  C18H23N5O3 357.4 18 23 5 3 0 0  0
## 650003 C18H18N4O3S 370.4 18 18 4 3 1 0  0
## 650004 C21H27N5O5S 461.5 21 27 5 5 1 0  0

Assign matrix data to data block:

 datablock(sdfset) <- propma 
## $`650001`
##           MF           MW            C            H            N 
## "C23H28N4O6"      "456.5"         "23"         "28"          "4" 
##            O            S            F           Cl 
##          "6"          "0"          "0"          "0"

String searching in SDFset ():

 grepSDFset("650001", sdfset, field="datablock", mode="subset") # Returns summary view of matches. Not printed here.
 grepSDFset("650001", sdfset, field="datablock", mode="index") 
## 1 1 1 1 1 1 1 1 1 
## 1 2 3 4 5 6 7 8 9

Export SDFset to SD file:

 write.SDF(sdfset[1:4], file="sub.sdf", sig=TRUE) 

Plot molecule structure of one or many SDFs:

 plot(sdfset[1:4], print=FALSE) # Plots structures to R graphics device 
plot of chunk plotstruct

plot of chunk plotstruct

 sdf.visualize(sdfset[1:4]) # Compound viewing in web browser 
Visualization webpage created by calling sdf.visualize.

Visualization webpage created by calling sdf.visualize.

Structure similarity searching and clustering:

 apset <- sdf2ap(sdfset) # Generate atom pair descriptor database for searching 
 data(apset) # Load sample apset data provided by library. 
 cmp.search(apset, apset[1], type=3, cutoff = 0.3, quiet=TRUE) # Search apset database with single compound. 
##   index    cid scores
## 1     1 650001 1.0000
## 2    96 650102 0.3517
## 3    67 650072 0.3118
## 4    88 650094 0.3095
## 5    15 650015 0.3011
 cmp.cluster(db=apset, cutoff = c(0.65, 0.5), quiet=TRUE)[1:4,] # Binning clustering using variable similarity cutoffs. 
## sorting result...
##       ids CLSZ_0.65 CLID_0.65 CLSZ_0.5 CLID_0.5
## 48 650049         2        48        2       48
## 49 650050         2        48        2       48
## 54 650059         2        54        2       54
## 55 650060         2        54        2       54

OpenBabel Functions

ChemmineR integrates now a subset of cheminformatics functionalities implemented in the OpenBabel C++ library (O’Boyle, Morley, and Hutchison, 2008; Cao, Charisi, Cheng, et al., 2008) . These utilities can be accessed by installing the ChemmineOB package and the OpenBabel software itself. ChemmineR will automatically detect the availability of ChemmineOB and make use of the additional utilities. The following lists the functions and methods that make use of OpenBabel. References are included to locate the sections in the manual where the utility and usage of these functions is described.

Structure format interconversions (see Section Format Inter-Conversions)

propOB: generates several compound properties. See the man page for a current list of properties computed.

## Note: the specification for class "RSWIGStruct" in package 'ChemmineOB' seems equivalent to one from package 'ChemmineR': not turning on duplicate class definitions for this class.
##                                                          cansmi
## 650001 O=C(Nc1noc(c1)C)CCC(=O)N(c1ccc2c(c1)OCCO2)CC(=O)NC1CCCC1
##                                                        cansmiNS    formula
## 650001 O=C(Nc1noc(c1)C)CCC(=O)N(c1ccc2c(c1)OCCO2)CC(=O)NC1CCCC1 C23H28N4O6
##         title HBA1 HBA2 HBD  logP    MR    MW nF TPSA
## 650001 650001   37   10   2 3.029 119.9 456.5  0  123

fingerprintOB: generates fingerprints for compounds. The fingerprint name can be anything supported by OpenBabel. See the man page for a current list.

## An instance of a 1024 bit "FPset" of type "FP2" with 100 molecules

smartsSearchOB: find matches of SMARTS patterns in compounds

#count rotable bonds
## 650001 650002 650003 650004 650005 
##     24     20     14     30     10

Overview of Classes and Functions

The following list gives an overview of the most important S4 classes, methods and functions available in the ChemmineR package. The help documents of the package provide much more detailed information on each utility. The standard R help documents for these utilities can be accessed with this syntax: ?function_name (e.g. ?cid) and ?class_name-class (e.g. ?“SDFset-class”).

Molecular Structure Data


Functions/Methods (mainly for SDFset container, SMIset should be coerced with smiles2sdf to SDFset)

Structure Descriptor Data



Import of Compounds

SDF Import

The following gives an overview of the most important import/export functionalities for small molecules provided by ChemmineR. The given example creates an instance of the SDFset class using as sample data set the first 100 compounds from this PubChem SD file (SDF): Compound_00650001_00675000.sdf.gz (ftp://ftp.ncbi.nih.gov/pubchem/Compound/CURRENT-Full/SDF/).

SDFs can be imported with the read.SDFset function:

 sdfset <- read.SDFset("http://faculty.ucr.edu/ tgirke/Documents/R_BioCond/Samples/sdfsample.sdf") 
 data(sdfsample) # Loads the same SDFset provided by the library 
 sdfset <- sdfsample
 valid <- validSDF(sdfset) # Identifies invalid SDFs in SDFset objects 
 sdfset <- sdfset[valid] # Removes invalid SDFs, if there are any 

Import SD file into SDFstr container:

r sdfstr <- read.SDFstr("http://faculty.ucr.edu/ tgirke/Documents/R_BioCond/Samples/sdfsample.sdf") Create SDFset from SDFstr class:

 sdfstr <- as(sdfset, "SDFstr") 
## An instance of "SDFstr" with 100 molecules
 as(sdfstr, "SDFset") 
## An instance of "SDFset" with 100 molecules


The read.SMIset function imports one or many molecules from a SMILES file and stores them in a SMIset container. The input file is expected to contain one SMILES string per row with tab-separated compound identifiers at the end of each line. The compound identifiers are optional.

Create sample SMILES file and then import it:

 data(smisample); smiset <- smisample
 write.SMI(smiset[1:4], file="sub.smi") 
 smiset <- read.SMIset("sub.smi")

Inspect content of SMIset:

 data(smisample) # Loads the same SMIset provided by the library 
 smiset <- smisample
## An instance of "SMIset" with 100 molecules
## $`650001`
## An instance of "SMI"
## [1] "O=C(NC1CCCC1)CN(c1cc2OCCOc2cc1)C(=O)CCC(=O)Nc1noc(c1)C"
## $`650002`
## An instance of "SMI"
## [1] "O=c1[nH]c(=O)n(c2nc(n(CCCc3ccccc3)c12)NCCCO)C"

Accessor functions:

## [1] "650001" "650002" "650003" "650004"
 smi <- as.character(smiset[1:2])

Create SMIset from named character vector:

 as(smi, "SMIset") 
## An instance of "SMIset" with 2 molecules

Export of Compounds

SDF Export

Write objects of classes SDFset/SDFstr/SDF to SD file:

 write.SDF(sdfset[1:4], file="sub.sdf") 

Writing customized SDFset to file containing ChemmineR signature, IDs from SDFset and no data block:

 write.SDF(sdfset[1:4], file="sub.sdf", sig=TRUE, cid=TRUE, db=NULL) 

Example for injecting a custom matrix/data frame into the data block of an SDFset and then writing it to an SD file:

 props <- data.frame(MF=MF(sdfset), MW=MW(sdfset), atomcountMA(sdfset)) 
 datablock(sdfset) <- props
 write.SDF(sdfset[1:4], file="sub.sdf", sig=TRUE, cid=TRUE) 

Indirect export via SDFstr object:

 sdf2str(sdf=sdfset[[1]], sig=TRUE, cid=TRUE) # Uses default components 
 sdf2str(sdf=sdfset[[1]], head=letters[1:4], db=NULL) # Uses custom components for header and data block 

Write SDF, SDFset or SDFstr classes to file:

 write.SDF(sdfset[1:4], file="sub.sdf", sig=TRUE, cid=TRUE, db=NULL)
 write.SDF(sdfstr[1:4], file="sub.sdf") 
 cat(unlist(as(sdfstr[1:4], "list")), file="sub.sdf", sep="") 


Write objects of class SMIset to SMILES file with and without compound identifiers:

 data(smisample); smiset <- smisample # Sample data set 

 write.SMI(smiset[1:4], file="sub.smi", cid=TRUE) write.SMI(smiset[1:4], file="sub.smi", cid=FALSE) 

Format Interconversions

The sdf2smiles and smiles2sdf functions provide format interconversion between SMILES strings (Simplified Molecular Input Line Entry Specification) and SDFset containers.

Convert an SDFset container to a SMILES character string:

 sdfset <- sdfsample[1] 
 smiles <- sdf2smiles(sdfset) 

Convert a SMILES character string to an SDFset container:

 sdf <- smiles2sdf("CC(=O)OC1=CC=CC=C1C(=O)O")

When the ChemineOB package is installed these conversions are performed with the OpenBabel Open Source Chemistry Toolbox. Otherwise the functions will fall back to using the ChemMine Tools web service for this operation. The latter will require internet connectivity and is limited to only the first compound given. ChemmineOB provides access to the compound format conversion functions of OpenBabel. Currently, over 160 formats are supported by OpenBabel. The functions convertFormat and convertFormatFile can be used to convert files or strings between any two formats supported by OpenBabel. For example, to convert a SMILES string to an SDF string, one can use the convertFormat function.

 sdfStr <- convertFormat("SMI","SDF","CC(=O)OC1=CC=CC=C1C(=O)O_name") 

This will return the given compound as an SDF formatted string. 2D coordinates are also computed and included in the resulting SDF string.

To convert a file with compounds encoded in one format to another format, the convertFormatFile function can be used instead.


To see the whole list of file formats supported by OpenBabel, one can run from the command-line “obabel -L formats”.

Splitting SD Files

The following write.SDFsplit function allows to split SD Files into any number of smaller SD Files. This can become important when working with very big SD Files. Users should note that this function can output many files, thus one should run it in a dedicated directory!

Create sample SD File with 100 molecules:

 write.SDF(sdfset, "test.sdf") 

Read in sample SD File. Note: reading file into SDFstr is much faster than into SDFset:

 sdfstr <- read.SDFstr("test.sdf") 

Run export on SDFstr object:

 write.SDFsplit(x=sdfstr, filetag="myfile", nmol=10) # 'nmol' defines the number of molecules to write to each file 

Run export on SDFset object:

 write.SDFsplit(x=sdfset, filetag="myfile", nmol=10) 

Streaming Through Large SD Files

The sdfStream function allows to stream through SD Files with millions of molecules without consuming much memory. During this process any set of descriptors, supported by ChemmineR, can be computed (e.g. atom pairs, molecular properties, etc.), as long as they can be returned in tabular format. In addition to descriptor values, the function returns a line index that gives the start and end positions of each molecule in the source SD File. This line index can be used by the downstream read.SDFindex function to retrieve specific molecules of interest from the source SD File without reading the entire file into R. The following outlines the typical workflow of this streaming functionality in ChemmineR.

Create sample SD File with 100 molecules:

 write.SDF(sdfset, "test.sdf") 

Define descriptor set in a simple function:

 desc <- function(sdfset) 
    # datablock2ma(datablocklist=datablock(sdfset)), 
    groups(sdfset), APFP=desc2fp(x=sdf2ap(sdfset), descnames=1024,
    type="character"), AP=sdf2ap(sdfset, type="character"), rings(sdfset,
    type="count", upper=6, arom=TRUE) )  

Run sdfStream with desc function and write results to a file called matrix.xls:

 sdfStream(input="test.sdf", output="matrix.xls", fct=desc, Nlines=1000) # 'Nlines': number of lines to read from input SD File at a time 

One can also start reading from a specific line number in the SD file. The following example starts at line number 950. This is useful for restarting and debugging the process. With append=TRUE the result can be appended to an existing file.

 sdfStream(input="test.sdf", output="matrix2.xls", append=FALSE, fct=desc, Nlines=1000, startline=950) 

Select molecules meeting certain property criteria from SD File using line index generated by previous sdfStream step:

 indexDF <- read.delim("matrix.xls", row.names=1)[,1:4] 
 indexDFsub <- indexDF[indexDF$MW < 400, ] # Selects molecules with MW < 400 
 sdfset <- read.SDFindex(file="test.sdf", index=indexDFsub, type="SDFset") # Collects results in 'SDFset' container 

Write results directly to SD file without storing larger numbers of molecules in memory:

 read.SDFindex(file="test.sdf", index=indexDFsub, type="file",

Read AP/APFP strings from file into APset or FP object:

 apset <- read.AP(x="matrix.xls", type="ap", colid="AP") 
 apfp <- read.AP(x="matrix.xls", type="fp", colid="APFP") 

Alternatively, one can provide the AP/APFP strings in a named character vector:

 apset <- read.AP(x=sdf2ap(sdfset[1:20], type="character"), type="ap") 
 fpchar <- desc2fp(sdf2ap(sdfset[1:20]), descnames=1024, type="character")
 fpset <- as(fpchar, "FPset") 

Storing Compounds in an SQL Database

As an alternative to sdfStream, there is now also an option to store data in an SQL database, which then allows for fast queries and compound retrieval. The default database is SQLite, but any other SQL database should work with some minor modifications to the table definitions, which are stored in schema/compounds.SQLite under the ChemmineR package directory. Compounds are stored in their entirety in the databases so there is no need to keep any original data files.

Users can define their own set of compound features to compute and store when loading new compounds. Each of these features will be stored in its own, indexed table. Searches can then be performed using these features to quickly find specific compounds. Compounds can always be retrieved quickly because of the database index, no need to scan a large compound file. In addition to user defined features, descriptors can also be computed and stored for each compound.

A new database can be created with the initDb function. This takes either an existing database connection, or a filename. If a filename is given then an SQLite database connection is created. It then ensures that the required tables exist and creates them if not. The connection object is then returned. This function can be called safely on the same connection or database many times and will not delete any data.

Loading Data

The functions loadSdf and loadSmiles can be used to load compound data from either a file (both) or an SDFset (loadSdf only). The fct parameter should be a function to extract features from the data. It will be handed an SDFset generated from the data being loaded. This may be done in batches, so there is no guarantee that the given SDFSset will contain the whole dataset. This function should return a data frame with a column for each feature and a row for each compound given. The order of the final data frame should be the same as that of the SDFset. The column names will become the feature names. Each of these features will become a new, indexed, table in the database which can be used later to search for compounds.

The descriptors parameter can be a function which computes descriptors. This function will also be given an SDFset object, which may be done in batches. It should return a data frame with the following two columns: “descriptor” and “descriptor_type”. The “descriptor” column should contain a string representation of the descriptor, and “descriptor_type” is the type of the descriptor. Our convention for atom pair is “ap” and “fp” for finger print. The order should also be maintained.

When the data has been loaded, loadSdf will return the compound id numbers of each compound loaded. These compound id numbers are computed by the database and are not extracted from the compound data itself. They can be used to quickly retrieve compounds later.

New features can also be added using this function. However, all compounds must have all features so if new features are added to a new set of compounds, all existing features must be computable by the fct function given. If new features are detected, all existing compounds will be run through fct in order to compute the new features for them as well.

For example, if dataset X is loaded with features F1 and F2, and then at a later time we load dataset Y with new feature F3, the fct function used to load dataset Y must compute and return features F1, F2, and F3. loadSdf will call fct with both datasets X and Y so that all features are available for all compounds. If any features are missing an error will be raised. If just new features are being added, but no new compounds, use the addNewFeatures function.

In this example, we create a new database called “test.db” and load it with data from an SDFset. We also define fct to compute the molecular weight, “MW”, and the number of rings and aromatic rings. The rings function actually returns a data frame with columns “RINGS” and “AROMATIC”, which will be merged into the data frame being created which will also contain the “MW” column. These will be the names used for these features and must be used when searching with them. Finally, the new compound ids are returned and stored in the “ids” variable.


 #create and initialize a new SQLite database 
 conn <- initDb("test.db")
## Loading required package: RSQLite
## Loading required package: DBI
 # load data and compute 3 features: molecular weight, with the MW function, 
 # and counts for RINGS and AROMATIC, as computed by rings, which 
 # returns a data frame itself. 
 ids<-loadSdf(conn,sdfsample, function(sdfset) 
 data.frame(MW = MW(sdfset), rings(sdfset,type="count",upper=6, arom=TRUE)) ) 


By default the loadSdf / loadSmiles functions will detect duplicate compound entries and only insert one of them. This means it is safe to run these functions on the same data set several times and you won’t end up with duplicates. This allows the functions to be re-run in the event that a previous run on a dataset does not complete. Duplicate compounds are detected by compouting the MD5 checksum on the textual representation of it.

It can also update existing compounds with new versions of the same compound. To enable this, set updateByName to true. It will then consider two compounds with the same name to be the same, even if the definition is different. Then, if the name of a compound exists in the database and it is trying to insert another compound with the same name, it will overwrite the existing compound. It will also drop and re-compute all associated descriptors and features for the new compound (assuming the required functions for descriptor and feature computation are available at the time the update is performed).

Duplicate Descriptors

It is often the case when loading a large set of compounds that several compounds will produce the same descriptor. ChemmineR detects this case and only stores one copy of the descriptor for every compound it is for. This feature saves some space and some time for processes that need to be applied to every descriptor. It also highlights a new problem. If you have a descriptor in hand and you want to find a single compound to represent it, which compound should be used if the descriptor was produced from multiple compounds? To address this problem, ChemmineR allows you to set priority values for each compound-descriptor mapping. Then, in contexts where a single compound is required, the highest priority compound will be chosen. Highest priority corresponds to the lowest numerical value. So mapping with priority 0 would be used first.

To set these priorities there is the function setPriorities. It takes a function, priorityFn, for computing these priority values. The setPriorities function should be run after loading a complete set of data. It will find each group of compounds which share the same descriptor and call the given function, priorityFn, with the compound_id numbers of the group. This function should then assign priorities to each compound-descriptor pair, however it wishes.

One built in priority function is forestSizePriorities. This simply prefers compounds with fewer disconnected components over compounds with more dissconnected components.



Compounds can be searched for using the findCompounds function. This function takes a connection object, a vector of feature names used in the tests, and finally, a vector of tests that must all pass for a compound to be included in the result set. Each test should be a boolean expression. For example: “c(”MW <= 400“,”RINGS > 3“)” would return all compounds with a molecular weight of 400 or less and more than 3 rings, assuming these features exist in the database. The syntax for each test is “<feature name> <SQL operator> <value>”. If you know SQL you can go beyond this basic syntax. These tests will simply be concatenated together with “AND” in-between them and tacked on the end of a WHERE clause of an SQL statement. So any SQL that will work in that context is fine. The function will return a list of compound ids, the actual compounds can be fetched with getCompounds. If just the names are needed, the getCompoundNames function can be used. Compounds can also be fetched by name using the findCompoundsByName function.

In this example we search for compounds with molecular weight less than 300. We then fetch the matching compounds and show their molecular weight.

 lightIds <- findCompounds(conn,"MW",c("MW < 300")) 
##   206   214   217   222   224   228   229   230   233   234   236   237 
## 262.3 240.3 265.3 287.4 266.7 278.3 282.3 270.3 223.2 265.3 249.2 285.3 
##   242   246   250   255   256   257   268   273   276   282   289   295 
## 229.3 248.3 266.3 275.4 299.8 263.4 296.3 275.4 294.4 140.1 294.4 276.3
 #names of matching compounds: 
##  [1] "650006" "650014" "650017" "650023" "650025" "650029" "650030"
##  [8] "650031" "650034" "650035" "650037" "650038" "650043" "650047"
## [15] "650052" "650060" "650061" "650062" "650073" "650078" "650081"
## [22] "650088" "650095" "650101"

Working with SDF/SDFset Classes

Several methods are available to return the different data components of SDF/SDFset containers in batches. The following examples list the most important ones. To save space their content is not printed in the manual.

 view(sdfset[1:4]) # Summary view of several molecules 

 length(sdfset) # Returns number of molecules 
 sdfset[[1]] # Returns single molecule from SDFset as SDF object 

 sdfset[[1]][[2]] # Returns atom block from first compound as matrix

 c(sdfset[1:4], sdfset[5:8]) # Concatenation of several SDFsets 

The grepSDFset function allows string matching/searching on the different data components in SDFset. By default the function returns a SDF summary of the matching entries. Alternatively, an index of the matches can be returned with the setting mode=“index”.

 grepSDFset("650001", sdfset, field="datablock", mode="subset") # To return index, set mode="index") 

Utilities to maintain unique compound IDs:

 sdfid(sdfset[1:4]) # Retrieves CMP IDs from Molecule Name field in header block. 
 cid(sdfset[1:4]) # Retrieves CMP IDs from ID slot in SDFset. 
 unique_ids <- makeUnique(sdfid(sdfset)) # Creates unique IDs by appending a counter to duplicates. 
 cid(sdfset) <- unique_ids # Assigns uniquified IDs to ID slot 

Subsetting by character, index and logical vectors:

 view(sdfset[c("650001", "650012")])
 mylog <- cid(sdfset)

Accessing SDF/SDFset components: header, atom, bond and data blocks:

 atomblock(sdf); sdf[[2]];
 sdf[["atomblock"]] # All three methods return the same component

 header(sdfset[1:4]) atomblock(sdfset[1:4])
 bondblock(sdfset[1:4]) datablock(sdfset[1:4]) header(sdfset[[1]])
 atomblock(sdfset[[1]]) bondblock(sdfset[[1]]) datablock(sdfset[[1]]) 

Replacement Methods:

 sdfset[[1]][[2]][1,1] <- 999 atomblock(sdfset)[1] <-
 atomblock(sdfset)[2] datablock(sdfset)[1] <- datablock(sdfset)[2] 

Assign matrix data to data block:

 datablock(sdfset) <- as.matrix(iris[1:100,])

Class coercions from SDFstr to list, SDF and SDFset:

 as(sdfstr[1:2], "list") as(sdfstr[[1]], "SDF")
 as(sdfstr[1:2], "SDFset") 

Class coercions from SDF to SDFstr, SDFset, list with SDF sub-components:

 sdfcomplist <- as(sdf, "list") sdfcomplist <-
 as(sdfset[1:4], "list"); as(sdfcomplist[[1]], "SDF") sdflist <-
 as(sdfset[1:4], "SDF"); as(sdflist, "SDFset") as(sdfset[[1]], "SDFstr")
 as(sdfset[[1]], "SDFset") 

Class coercions from SDFset to lists with components consisting of SDF or sub-components:

 as(sdfset[1:4], "SDF") as(sdfset[1:4], "list") as(sdfset[1:4], "SDFstr")

Molecular Property Functions (Physicochemical Descriptors)

Several methods and functions are available to compute basic compound descriptors, such as molecular formula (MF), molecular weight (MW), and frequencies of atoms and functional groups. In many of these functions, it is important to set addH=TRUE in order to include/add hydrogens that are often not specified in an SD file.

 propma <- atomcountMA(sdfset, addH=FALSE) 
 boxplot(propma, col="blue", main="Atom Frequency") 
plot of chunk boxplot

plot of chunk boxplot

 boxplot(rowSums(propma), main="All Atom Frequency") 

Data frame provided by library containing atom names, atom symbols, standard atomic weights, group and period numbers:

##   Number      Name Symbol Atomic_weight Group Period
## 1      1  hydrogen      H         1.008     1      1
## 2      2    helium     He         4.003    18      1
## 3      3   lithium     Li         6.941     1      2
## 4      4 beryllium     Be         9.012     2      2

Compute MW and formula:

 MW(sdfset[1:4], addH=FALSE)
##  CMP1  CMP2  CMP3  CMP4 
## 456.5 357.4 370.4 461.5
 MF(sdfset[1:4], addH=FALSE) 
##          CMP1          CMP2          CMP3          CMP4 
##  "C23H28N4O6"  "C18H23N5O3" "C18H18N4O3S" "C21H27N5O5S"

Enumerate functional groups:

 groups(sdfset[1:4], groups="fctgroup", type="countMA") 
## CMP1    0    2   1     0   0    0    0     0     0   2    0   0
## CMP2    0    2   2     0   1    0    0     0     0   0    0   0
## CMP3    0    1   1     0   1    0    1     0     0   0    0   0
## CMP4    0    1   3     0   0    0    0     0     0   2    0   0

Combine MW, MF, charges, atom counts, functional group counts and ring counts in one data frame:

 propma <- data.frame(MF=MF(sdfset, addH=FALSE), MW=MW(sdfset, addH=FALSE),
                             Ncharges=sapply(bonds(sdfset, type="charge"), length),
                             atomcountMA(sdfset, addH=FALSE), 
                             groups(sdfset, type="countMA"), 
                             rings(sdfset, upper=6, type="count", arom=TRUE))
##               MF    MW Ncharges  C  H N O S F Cl RNH2 R2NH R3N ROPO3 ROH
## CMP1  C23H28N4O6 456.5        0 23 28 4 6 0 0  0    0    2   1     0   0
## CMP2  C18H23N5O3 357.4        0 18 23 5 3 0 0  0    0    2   2     0   1
## CMP3 C18H18N4O3S 370.4        0 18 18 4 3 1 0  0    0    1   1     0   1
## CMP4 C21H27N5O5S 461.5        0 21 27 5 5 1 0  0    0    1   3     0   0
## CMP1    0    0     0     0   2    0   0     4        2
## CMP2    0    0     0     0   0    0   0     3        3
## CMP3    0    1     0     0   0    0   0     4        2
## CMP4    0    0     0     0   2    0   0     3        3

The following shows an example for assigning the values stored in a matrix (e.g. property descriptors) to the data block components in an SDFset. Each matrix row will be assigned to the corresponding slot position in the SDFset.

 datablock(sdfset) <- propma # Works with all SDF components 
 test <- apply(propma[1:4,], 1, function(x) 
 data.frame(col=colnames(propma), value=x)) 
 sdf.visualize(sdfset[1:4], extra = test) 

The data blocks in SDFs contain often important annotation information about compounds. The datablock2ma function returns this information as matrix for all compounds stored in an SDFset container. The splitNumChar function can then be used to organize all numeric columns in a numeric matrix and the character columns in a character matrix as components of a list object.

 datablocktag(sdfset, tag="PUBCHEM_NIST_INCHI")

Convert entire data block to matrix:

 blockmatrix <- datablock2ma(datablocklist=datablock(sdfset)) # Converts data block to matrix 
 numchar <- splitNumChar(blockmatrix=blockmatrix) # Splits matrix to numeric matrix and character matrix 
 numchar[[1]][1:4,]; numchar[[2]][1:4,]
 # Splits matrix to numeric matrix and character matrix 

Bond Matrices

Bond matrices provide an efficient data structure for many basic computations on small molecules. The function conMA creates this data structure from SDF and SDFset objects. The resulting bond matrix contains the atom labels in the row/column titles and the bond types in the data part. The labels are defined as follows: 0 is no connection, 1 is a single bond, 2 is a double bond and 3 is a triple bond.

 exclude=c("H")) # Create bond matrix for first two molecules in sdfset

 conMA(sdfset[[1]], exclude=c("H")) # Return bond matrix for first molecule 
 plot(sdfset[1], atomnum = TRUE, noHbonds=FALSE , no_print_atoms = "", atomcex=0.8) # Plot its structure with atom numbering 
 rowSums(conMA(sdfset[[1]], exclude=c("H"))) # Return number of non-H bonds for each atom

Charges and Missing Hydrogens

The function bonds returns information about the number of bonds, charges and missing hydrogens in SDF and SDFset objects. It is used by many other functions (e.g. MW, MF, atomcount, atomcuntMA and plot) to correct for missing hydrogens that are often not specified in SD files.

 bonds(sdfset[[1]], type="bonds")[1:4,]
##   atom Nbondcount Nbondrule charge
## 1    O          2         2      0
## 2    O          2         2      0
## 3    O          2         2      0
## 4    O          2         2      0
 bonds(sdfset[1:2], type="charge")
## $CMP1
## $CMP2
 bonds(sdfset[1:2], type="addNH") 
## CMP1 CMP2 
##    0    0

Ring Perception and Aromaticity Assignment

The function rings identifies all possible rings in one or many molecules (here sdfset[1]) using the exhaustive ring perception algorithm from Hanser, Jauffret, and Kaufmann (1996) . In addition, the function can return all smallest possible rings as well as aromaticity information.

The following example returns all possible rings in a list. The argument upper allows to specify an upper length limit for rings. Choosing smaller length limits will reduce the search space resulting in shortened compute times. Note: each ring is represented by a character vector of atom symbols that are numbered by their position in the atom block of the corresponding SDF/SDFset object.

 ringatoms <- rings(sdfset[1], upper=Inf, type="all", arom=FALSE, inner=FALSE)

For visual inspection, the corresponding compound structure can be plotted with the ring bonds highlighted in color:

 atomindex <- as.numeric(gsub(".*_", "", unique(unlist(ringatoms))))
 plot(sdfset[1], print=FALSE, colbonds=atomindex) 
plot of chunk unnamed-chunk-80

plot of chunk unnamed-chunk-80

Alternatively, one can include the atom numbers in the plot:

 plot(sdfset[1], print=FALSE, atomnum=TRUE, no_print_atoms="H") 
plot of chunk unnamed-chunk-81

plot of chunk unnamed-chunk-81

Aromaticity information of the rings can be returned in a logical vector by setting arom=TRUE:

 rings(sdfset[1], upper=Inf, type="all", arom=TRUE, inner=FALSE) 
## $RINGS$ring1
## [1] "N_10" "O_6"  "C_32" "C_31" "C_30"
## $RINGS$ring2
## [1] "C_12" "C_14" "C_15" "C_13" "C_11"
## $RINGS$ring3
## [1] "C_23" "O_2"  "C_27" "C_28" "O_3"  "C_25"
## $RINGS$ring4
## [1] "C_23" "C_21" "C_18" "C_22" "C_26" "C_25"
## $RINGS$ring5
##  [1] "O_3"  "C_28" "C_27" "O_2"  "C_23" "C_21" "C_18" "C_22" "C_26" "C_25"
## ring1 ring2 ring3 ring4 ring5 

Return rings with no more than 6 atoms that are also aromatic:

 rings(sdfset[1], upper=6, type="arom", arom=TRUE, inner=FALSE) 
## [1] "N_10" "O_6"  "C_32" "C_31" "C_30"
## [1] "C_23" "C_21" "C_18" "C_22" "C_26" "C_25"

Count shortest possible rings and their aromaticity assignments by setting type=count and inner=TRUE. The inner (smallest possible) rings are identified by first computing all possible rings and then selecting only the inner rings. For more details, consult the help documentation with ?rings.

 rings(sdfset[1:4], upper=Inf, type="count", arom=TRUE, inner=TRUE) 
## CMP1     4        2
## CMP2     3        3
## CMP3     4        2
## CMP4     3        3

Rendering Chemical Structure Images

R Graphics Device

A new plotting function for compound structures has been added to the package recently. This function uses the native R graphics device for generating compound depictions. At this point this function is still in an experimental developmental stage but should become stable soon.

If you have ChemmineOB available you can use the regenCoords option to have OpenBabel regenerate the coordinates for the compound. This can sometimes produce better looking plots.

Plot compound Structures with R’s graphics device:

 sdfset <- sdfsample
 plot(sdfset[1:4], regenCoords=TRUE,print=FALSE) # 'print=TRUE' returns SDF summaries
plot of chunk plotstruct2

plot of chunk plotstruct2

Customized plots:

 plot(sdfset[1:4], griddim=c(2,2), print_cid=letters[1:4], print=FALSE, 

In the following plot, the atom block position numbers in the SDF are printed next to the atom symbols (atomnum = TRUE). For more details, consult help documentation with ?plotStruc or ?plot.

 plot(sdfset["CMP1"], atomnum = TRUE, noHbonds=F , no_print_atoms = "",
        atomcex=0.8, sub=paste("MW:", MW(sdfsample["CMP1"])), print=FALSE) 
plot of chunk plotstruct3

plot of chunk plotstruct3

Substructure highlighting by atom numbers:

 plot(sdfset[1], print=FALSE, colbonds=c(22,26,25,3,28,27,2,23,21,18,8,19,20,24)) 
plot of chunk plotstruct4

plot of chunk plotstruct4

Online with ChemMine Tools

Alternatively, one can visualize compound structures with a standard web browser using the online ChemMine Tools service. The service allows to display other information next to the structures using the extra argument of the sdf.visualize function. The following examples demonstrate, how one can plot and annotate structures by passing on extra data as vector of character strings, matrices or lists.

Plot structures using web service ChemMine Tools:

Visualization webpage created by calling sdf.visualize.

Visualization webpage created by calling sdf.visualize.

Add extra annotation as v