BROOD FAQ
How do I get the BROOD fragment database?
The BROOD 2.0 fragment database is available here, as a .tar.gz “tarball”. Download the file brood-2.0.0-database.tar.gz and unpack it in the same location as the BROOD 2.0 app. Please note that this is a very large file (>3Gb) so may take a long time to download. The file is not OS specific so if you have problems it is possible to download on a different OS and then transfer it. Internet Explorer on Windows cannot cope with a file >2Gb so we recommend using a different browser e.g. Firefox.
During BROOD database construction, CHOMP fragments molecules to generate the final database. I have my own fragment library. How will CHOMP deal with the fragments during database generation?
Under the default behavior, CHOMP takes a whole molecule (rather than fragments) as input, breaks them apart, filters them, collects a unique set, generates conformers, then finally, generates the BROOD database (including pre-calculations for the search.
If you are starting with fragments rather than whole molecules. pass the fragments directly to CHOMP using
the "-userFrags" flag. Those fragments will go through the final 2 steps of the CHOMP process; conformer generation and database construction. Please note that fragments specified by "-userFrags" will still go through duplicate removal relative to the default BROOD database (this can be turned off with "-removeDuplicates false") and relative to any fragments specified with the "-userUnique" flag.
Is the name of the molecule or Id number saved with the fragments in the database created by CHOMP?
In the current release of CHOMP (version 2.0), there is not a 1 to 1 mapping of molecules to fragments.
chomp -in input.sdf -out output.ism -omega false
and when I inspected output.ism, I noticed that all chiral information had been removed. How does CHOMP handle stereochemistry?
In the first part of database generation, CHOMP cuts apart a known molecule to get fragments with known chemistry. The intention is to collect fragments without stereo and then enumerate the stereo to give all feasible fragments with the acceptable chemistry. This is why the intermediate .ism file does not contain stereo. Stereochemistry is enumerated in the second part of database generation.
chomp -userFrags output.ism -out DB
I noticed the line "Number of Warnings = 100" in the DB.info file, but I cannot find this information in the log file. How can I get detailed information about these warnings?
If there are no warning messages in the log file, there is nothing to worry about. CHOMP suppresses some minor warnings that are not important in the context in which CHOMP is using a function.
Does CHOMP assign fragment partial charges before or after fragmentation?
Charges are assigned after fragmentation.
Many scores are displayed in the BROOD result spreadsheet. Are these scores calculated between the query fragment and the fragment in the database or are they calculated between the whole query molecule and the whole molecule of the result?
All the scores are fragment-fragment scores. The only score associated with the whole molecule is the pActive Abbott model, which is only defined for whole molecules.
What charge model is used for assigning partial charges in BROOD?
MMFF partial charges model. For electrostatic similarity, OpenEye has explored several charge sets. While improved partial charges such as AM1-BCC, ESP, and RESP are better for physical properties, for electrostatic similarity this is not the case. Instead, consistent, conformer-independent partial charges are most important and thus BROOD uses MMFF partial charges.
Does BROOD re-assign charges on the result structures?
Yes, charges are cleared and then set to MMFF partial charges prior to the MMFF optimization of the built structures.
How is pH calculated in BROOD?
pH is not calculated. New molecules are set to a protonation state of pH~7.4 using a simple rule-based system. This does not take into account the protein environment, which will certainly effect some cases.
In the BROOD setup screen, there is an Advanced Search setting for "Min. shape Tanimoto". At what stage of the search is this score calculated: before fragment replacement, after fragment replacement but before minimization, or after fragment replacement and minimization?
Minimum Shape Tanimoto is calculated before fragment replacement after (fragment) minimization. It is an advanced setting that can be useful if you don't get any hits back from a search or have other specialized needs. However, it may have unexpected and/or subtle side effects. In many cases, exploring the "link Only" option can be more productive.
Does the default BROOD database contain cation or anion fragments?
Yes.
Can a user have a cation/anion in a query?
Yes, cations and anions in a query are acceptable.
How does BROOD and CHOMP handle ions?
BROOD and CHOMP carefully track ionization states and maintain them up until a fragment is built into a new molecule. At this point new bonds are formed, so ionization is reassessed. This reassessment can be turned off if desired.
If a query has an anion, do the result molecules contain the same anion and/or have its neutralized form?
If a query has an anion, the returned hits (as fragments) will be assessed separately (if anion and neutral). When built into a whole molecule, the anion can be reset to a pH=7 form (by default), or left as is.
Does CHOMP use OMEGA defaults when generating 3D structures of the fragments?
No, CHOMP doesn’t use the defaults for OMEGA, which are designed for whole molecules. CHOMP is specifically tuned for fragments using OMEGA and the false negatives from the Lommerse test set (M. Wagener and J. P. M. Lommerse, J. Chem. Inf. Model. 2006, 46, 677 - 685) to develop new OMEGA parameters for use with fragments in CHOMP. See also http://www.eyesopen.com/2007_cup_presentations/CUP8_Skillman.pdf.