WolframChemistry/IsomerGeneration

(0.0.9) current version: 1.0.3 »

Generate Isomers from Chemical Formulas

Contributed By: Jason Biggs, Wolfram Research

Starting from a chemical formula, generate all possible molecular structures with correct valence. Quickly create millions of isomers or more and read them in as "SMILES" strings or Molecule objects using a stream-like interface. Customize the structure generation by filtering out predefined substructure patterns, e.g. triple bonds in small rings or molecules with bridging double bonds, or define new valence rules for atoms outside the common organic set. This package uses the external application surge, based on the canonical generation path method.

Installation Instructions

To install this paclet in your Wolfram Language environment, evaluate this code:
PacletInstall["WolframChemistry/IsomerGeneration"]

Details

The functions defined in the IsomerGeneration paclet rely on the open-source application Surge to generate chemical structures. See "Surge: A Fast Open-Source Chemical Graph Generator" for implementation details.

Options are provided to pre-filter certain substructures. Specify the number of bonds, only create isomers with 2 or fewer 5-membered rings, require the underlying Graph to be PlanarGraphQ, exclude molecules with triple bonds, etc.

Elements supported include carbon, boron, nitrogen, phosphorus, oxygen, sulfur, hydrogen, chlorine, fluorine, bromine, and iodine at their lowest valences. Other elements can be included by providing their valence.

Ions (charged molecules) and zwitterions (neutral molecules with positive and negative parts) are not supported.

Stereoisomers are not generated.

For many chemical formulas it is possible to generate billions of isomers, with memory footprints of many gigabytes. The isomers are stored as text files filled with "SMILES" strings, located easily in your $WolframDocumentsDirectory.

Paclet Guide

Examples

Basic Examples (2)

List and plot all the isomers of dichlorobutane:

In[1]:=

Out[1]=

In[2]:=

Out[2]=

Count the number of isomers of C₁₀H₁₆O without listing them:

In[3]:=

Out[3]=

Substitute fluorine for two of they hydrogens and count the isomers:

In[4]:=

Out[4]=

Generate all the isomers for C10H16OF2 asynchronously:

In[5]:=

Out[5]=

Check the status of the isomer generation:

In[6]:=

Out[6]=

Find information about the isomers:

In[7]:=

Out[7]=

Read the first SMILES string from the result:

In[8]:=

Out[8]=

Read the next entry as a Molecule:

In[9]:=

Out[9]=

Search through all 3.5 million isomers, lazily selecting only the first 100 that have both fluorine atoms on the same carbon atom:

In[10]:=

Out[10]=

In[11]:=

Out[11]=

Scope (6)

The system function FindIsomers will find only find isomers that have been submitted to the "PubChem" database. This can be useful because it biases the results towards chemicals that have been found in nature or have been successfully synthesized, but it doesn't allow discovery of wholly new compounds. Take this medium-sized organic compound as an example:

In[12]:=