This directory represents a json snapshot of the PL Signal Transduction
Model (STM) knowldege bases and generated models as of 16nov28.

The json representation consists of 2016datums.json   (over 58k datums, 56mb) and five folders (also as zips): JSONB, cltJSONU, DishNets, cltJSONUnc, DishNetsnc explained below.   

It is assumed that the user is familiar with json and 
the PL concepts.  More information available on request to clt@csl.sri.com

DATUMS 

Datums are a computable representation of experimental results, used
to infer rules in the STM models. 16dec28datums.json is a slightly
later export of the datum collection (57mb, over 60k datums).
datum-json-scheme.txt is an informal description of the structure of
datum objects. See pl.csl.sri.com/datumkb.html for a description of
datum elements.


JSONB/ 
This folder is a representation of the full controlled vocabulary
 basicops.json 
    --- maps PL names to descriptions, including metadata with fields 
    --- for reference database accession and synonyms  
    --- includes all of the protein, chemical, ... ops curated
    --- whether or not they appear in a rule  
                                
 sorts.json     --- the sort hierarchy for the executable model
 datumsorts.json  --- the sort hierarchy for datum parsing
                  --- includes the above, and things like handles
                  --- assays, detection ops, ....

The rule kb comes in two flavors, one generated included common rules
(rule label begins with digits followed by "c." vs digits followd by
".") and one generated without including common rules. The former is
much bigger. Dishes represent experimental setups -- what cells are
expressing and any treatment. Dishnets add reachable rules to the
dish making an executable model. In some dishes common rules are
essential to fill gaps. In some cases, like the EgfDish, which is
more highly curated, they tend to add noise. Common rules may be
useful in starting with a model generated with out them and more
selectively adding those that improve the model instead of the PLA
simple automatic add them all approach.

The json for the RKB without common rules is contained in
  cltJSONUnc which contains
     basicOps.json	--- as above restricted to entities used
     locs.json	--- descriptions of locations
     mods.json	--- descriptions of modifications
     occs.json	--- descriptions of occurrences 
          --- refers (by name) to entities in basicOps, locs and mods
     rules.json 
       --- each rule has a name/identifier, and three lists 
       --- of occurrence names: controls, consumed, produced
       --- consumed and produced are the rule input (left hand side)
       --- and output (righthand side).  Controls represents the
       --- required biological context such as enzymes, cofactors ...
       --- Each rule also has an evidence field
and
  DishNetsnc --- json for each dish (experimental setup) and its reachable rules.
   --- each dishnet has fields init (named of the occurrences in the 
   --- initial state), and rids (names of the rules).  The names of 
   --- occurrences and rules refer to the desctipions in occs.json
   --- and rules.json

The json for the RKB with common rules is contained in
  cltJSONU --- the KB 
and
  DishNets --- json for each dish (experimental setup)

with contents analogous to the above.


Some statistics

STM7
  rules  1575
  occs   1378
  EgfDish 215 rules, 127 initOccs
  
STM with Common Rules  
  rules   5220
  occs    2216
  EgfDish 904 rules, 144 init occs

STM with out Common Rules  
  rules  2131
  occs   1783
  Egfdish 297 rules 150 init occs