ICRXN: The reaction processing package

Do you have millions of reaction data in your company’s Electronic Lab Notebook that might be:

 
  • Erroneous

  • Incomplete

  • Duplicated

  • Unstructured

  • Non-normalised

ICRXN is a processing package combining four components that will help scientists and cheminformaticians to 

 
  • Identify

  • Clean

  • Organise

  • Structure

  • Cluster

  • Analyze

  • Augment

  • Correct

huge amounts of chemical reaction data in electronic format (RXN file, RD file or all other reactions formats which can be converted into these). As a result, you can identify synthetically valuable reactions.

The ICRXN bundle consists of four leading reaction processing and cleaning tools developed by InfoChem using their extensive knowledge and expertise:

ICMAP: Identify - applying AI this simulates a chemical expert using our proven market leading algorithm for atom-atom mapping. Atom-atom mapping of reactions constitutes essential information in computational handling of chemical reactions such as (substructure or similarity) searching, reaction center identification, reaction classification and indexing.

CLASSIFY: Cluster, Organize, Analyze - The CLASSIFY algorithm calculates topology-based reaction class  codes for given reactions in electronic format. It is used to calculate formal similarity of reactions and systematic clustering of reactions in databases. In retrieval systems it is used for post-search management of large hit sets and linking of reaction information from different sources. It enables quality control and overlap studies of reactions.

ICNameRXN: Augment, Cluster - As a specific sub functionality of CLASSIFY, ICNameRXN automatically assigns name reactions (such as aldol condensation, pinacol rearrangement or dials-alder reaction) to given reactions in electronic format. Besides much simplified reaction retrieval by name instead of complex reaction substructure searching, name reaction assignment is used to characterize, cluster and manage reaction databases. Named reactions are grouped in an intuitive hierarchical chemical reaction taxonomy.

ICRXNFILTER: Clean, Correct - This is used to filter and clean up large reaction databases. It is built on a series of heuristic and machine learning based algorithms to identify erroneous, incomplete or doubtful reactions.

All components are available as shared libraries for Linux or Windows dll’s but could also be offered in form of REST API’s if required.

For more information please contact:

Dr. Josef Eiblmaier

Tel            +49(0)89/2153716-12

Email        je@infochem.de

Or fill in our contact form here: https://www.infochem.de/contact