On Apr 8, 2019, at 19:59, Amber Alley <ara...@georgetown.edu> wrote: > > Does openbabel have the capability to produce fingerprints that are binary > only?
Yes. The MACCS, FP2, FP3, and FP4 fingerprints are fixed-length binary fingerprints. Open Babel also implements ECFP fingerprints, which are not binary fingerprints. > I ask this because I want to compare openbabel fingerprints with other > fingerprint methods that use binary fingerprints. You might look at my package, chemfp. The Python API includes features which make it easier to compare fingerprints from different sources. The URL is http://chemfp.com/ which links to the documentation. Version 1.5 is free software, and is also available through the Python package index. You can install it, for example, with: python -m pip install chemfp > If not, can I compare openbabel fingerprints (which are not binary) to > fingerprints that are binary using the Tanimoto coefficient? In general it is not meaningful to do that. For example, it would not be meaningful to compare the Tanimoto of (for example) an ECFP4 fingerprint, no matter how it is reduced to a binary string, to a Daylight-style linear hash fingerprint, even if they are the same length. The one exception I can think of is if you want to compare two fingerprints which are essentially the same type though from different implementation. For example, one might generate the MACCS key fingerprints from Open Babel and compare them to the MACCS key fingerprints generated from RDKit. This might be used to characterize the effect of different chemical perception algorithms on the resulting fingerprints. Cheers, Andrew da...@dalkescientific.com _______________________________________________ OpenBabel-discuss mailing list OpenBabel-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openbabel-discuss
