[OTDev] OpenTox Month 18 deliverables

Barry Hardy barry.hardy at douglasconnect.com
Tue Dec 22 13:35:48 CET 2009


Dear All:

For our present discussion and iteration 3/4 planning, I thought it 
would be useful to remind ourselves of our collective deliverables for 
Month 18 (end of February):

2.2    Prototype framework    ALU-FR    P    PU    Month 18

   1. Implementation of a prototype framework with APIs for work
      packages 3 (databases), 4 (algorithms) and 5 (validation)
      (Deliverable 2)
   2. Implementation of a prototype GUI for toxicological risk assessors
      (Deliverable 2)
   3. Implementation of a prototype GUI for model developers (Deliverable 2)

3.2    Prototype database    IDEA    P    PU    Month 18

   1. Definition of requirements for data inclusion (Deliverable 2)
   2. Identification of suitable data sources (Deliverable 2)
   3. Implementation of a prototype database according to the
      requirements from WP1 (Deliverable 2)
   4. Import of initial data into the prototype database (Deliverable 2)

(Note: I note that in the WP3 description this deliverable is mistakenly 
indicated as 12 Months unlike in the Deliverables Table)

4.2    Prototype algorithms   NTUA    P    PU    Month 18

   1. Integration and implementation of algorithms for the generation of
      structural features (e.g. paths, trees, subgraphs, multiple
      neighborhood of atoms, pharmacophore descriptors)
   2. Integration and implementation of algorithms for the calculation
      of chemical properties (e.g. logP, surface parameters, reactivity
      indices)
   3. Implementation of algorithms for the retrieval of bioassay data
      from sources like PubChem
   4. Integration and implementation of algorithms for feature selection
      (e.g. statistical filters, closed sets, principal component analysis)
   5. Integration and implementation of algorithms for classification
      and regression (e.g. k-nearest neighbors, linear regression,
      neural nets, support vector machines, decision and regression trees)
   6. Integration and implementation of algorithms for the aggregation
      of predictions (rule based and data driven)
   7. Integration and implementation of supporting algorithms (e.g.
      applicability domain estimation, various measures of chemical
      similarity, structure and property based searches)
   8. Implementation and a plugin system for external (commercial) programs

(Note: We did not specify which algorithms would be implemented for the 
18 Month Prototype

5.1    Prototype validation routines    ALU-FR   P    PU    Month 18

   1. Implementation of validation methods based on artificial test sets
      (e.g. cross-validation, leave-one-out, simple training/test set
      splits) (Deliverable 1)

Barry


More information about the Development mailing list