jCompoundMapper: A Library for Chemical Fingerprints
Author Information
Author(s): Hinselmann Georg, Rosenbaum Lars, Jahn Andreas, Fechner Nikolas, Zell Andreas
Primary Institution: University of Tübingen, Center for Bioinformatics Tübingen (ZBIT)
Hypothesis
The aim of this work is to introduce a library for exactly defined molecular decompositions, with a strong focus on the application of these features in machine learning and data mining.
Conclusion
jCompoundMapper is a library for chemical graph fingerprints that provides several options for data mining applications.
Supporting Evidence
- The library provides a command-line interface and can be used as a Java API.
- jCompoundMapper features a variety of fingerprinting algorithms for chemical graphs.
- The quality of the data mining results was tested using standard benchmarks.
Takeaway
jCompoundMapper is a tool that helps scientists understand and compare chemical compounds by creating unique fingerprints for them, which can be used in machine learning.
Methodology
The library implements various fingerprinting algorithms and allows for parameterization of encodings, including search depth and distance cut-offs.
Limitations
The performance of some encodings was slower compared to others, particularly those relying on pharmacophore point definitions.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website