combinatorial_synthesis Command Help
Command: $SCHRODINGER/utilities/combinatorial_synthesis
usage: combinatorial_synthesis [-h] [-NJOBS NJOBS]
[-HOST <hostname>]
[-JOBNAME JOBNAME]
[-out OUT]
[-print_route]
[-plot_route]
[-list_descriptors]
[-list_reactions]
[-print_reactions_table]
[-seed INT]
[-max_products MAX_PRODUCTS]
[-max_tries MAX_TRIES]
[-start START]
[-stop STOP]
[-RESTART STATE_FILENAME]
[-reactions_file REACTIONS_FILE]
[-reagent_source REAGENT=SOURCE]
[-debug]
[-id_property ID_PROPERTY]
[-descriptors DESCRIPTORS]
[-reagent_property_filter FILENAME]
[-product_property_filter FILENAME]
[-reagent_smarts_filter FILENAME]
[-product_smarts_filter FILENAME]
[-rpf REAGENT=FILENAME]
[-rsf REAGENT=FILENAME]
[-ref_mols FILENAME]
[-pre_filter]
[-filter_single_reagents]
[-no3d]
[-require_stereo]
[-v3000] [-cxsmiles]
[-no_multiple_products]
[-no_dedup] [-sort]
[-count_unique]
[-ncycles <N>]
[-write_failed_reactions]
[-systematic]
[-libpath LIBPATH]
[-sort_by_index_sum]
[route]
Enumerate all the products obtained from a synthetic route.
positional arguments:
route reaction name or JSON route file
options:
-h, --help show this help message and exit
-out OUT, -o OUT output file
-print_route print the route (do not enumerate)
-plot_route plot the route into PNG file (do not enumerate)
-list_descriptors list the names of all available RDKit descriptors and
exit
-list_reactions list the names of all available reactions
-print_reactions_table
print a table of reaction names along with
descriptions and reagent classes
-seed INT seed for random number generator (default: 42). If
zero, a random seed will be generated automatically.
-max_products MAX_PRODUCTS, -max MAX_PRODUCTS
maximum number of products to report (default: 10000);
use zero for unlimited
-max_tries MAX_TRIES maximum number of reactant combinations to try before
giving up (default is up to 100*max_products, but may
be less depending on the number of possible
combinations).
-start START reagent combination index to start enumeration from
for systematic enumeration. By default, starts from
first combination in the Cartesian product of all
reagent sources.
-stop STOP for systematic enumeration, specify the index at which
to stop (exclusive). The enumeration will process
combinations from start up to but not including stop.
Default: continue until max_products is reached or all
combinations are exhausted.
-RESTART STATE_FILENAME
use STATE_FILENAME to restart a failed job
-reactions_file REACTIONS_FILE
non-default JSON reactions file
-reagent_source REAGENT=SOURCE, -r REAGENT=SOURCE
use structures from SOURCE file for a given REAGENT,
identified by number
-debug very verbose mode
-id_property ID_PROPERTY
reagent id property to propagate to the product's
s_reaction_reagent_<n> properties (default: s_m_title)
-descriptors DESCRIPTORS
RDKit descriptors to compute for the products (comma-
separated list; default: MolLogP,MolWt,NumChiralCenter
s,NumHAcceptors,NumHDonors,TPSA)
-no3d Skip 3D coordinate generation
-require_stereo disallow generation of 3D structures with unspecified
stereochemistry
-v3000 use V3000 when writing SD files
-cxsmiles use CXSMILES when writing CSV files
-no_multiple_products
reject a synthesis when a reaction yields more than
one product
-no_dedup don't exclude duplicate products
-sort sort products by SMILES
-count_unique don't count duplicate products toward -max_products
(if there are any duplicates, more than -max_products
will be written out);
-ncycles <N> try to apply each reaction up to <N> times, to
generate polysubstituted products when a reaction can
be applied more than once (default is to use what is
specified for each reaction in the route file, or else
1)
-write_failed_reactions
write failed reactions out to
<jobname>_failed_reactions.csv for debugging
-systematic use systematic enumeration instead of random sampling
-libpath LIBPATH, -L LIBPATH
path to directory containing reagent files (may be
used more than once)
-sort_by_index_sum perform systematic enumeration in the order of the sum
of reagent indices within the source files rather than
in the order of Cartesian product of the sources.
Standard Options:
-NJOBS NJOBS Divide the overall job into NJOBS subjobs.
Job Control Options:
-HOST <hostname> Run job remotely on the indicated host entry.
-JOBNAME JOBNAME Provide an explicit name for the job.
Filtering Options:
-reagent_property_filter FILENAME
propfilter or JSON file to be applied to each reagent
source
-product_property_filter FILENAME
propfilter or JSON file to be applied to each product
-reagent_smarts_filter FILENAME
canvasSearch filter file to be applied to each reagent
source
-product_smarts_filter FILENAME
canvasSearch filter file to be applied to each product
-rpf REAGENT=FILENAME
propfilter or JSON file to be applied to an individual
reagent source
-rsf REAGENT=FILENAME
canvasSearch filter file to be applied to an
individual reagent source
-ref_mols FILENAME reference molecules for RDKit fingerprint-based
similarity calculations
-pre_filter apply reagent filters before enumerating (this is
always done for systematic enumeration)
-filter_single_reagents
apply reagent filters even to sources with only one
compound