Documentation for Mirela Andronescu's PhD thesis
"Computational approaches for RNA energy parameter estimation"
- UBC Computer Science 2008 -
- Supervisors: Anne Condon and Holger Hoos -

• Thesis [ PDF file , PPT presentation, TEX files, all sources ]
  
• Parameter sets from Table 5.8
  • Description of parameters [ TXT file ]
  • BL* [ TXT file ]
  • BL (no dangles) [ TXT file ]
  • LAM-CG (CG*) [ TXT file ]
  • NOM-CG [ TXT file ]
  • DIM-CG [ TXT file ]
    
  • Turner99 (363 parameters) [ TXT file ]
  • Turner99 (315 parameters, i.e. 363 parameters with dangling ends = 0) [ TXT file ]
• Data sets
  • The RNA STRAND v2.0 database [ link to the online database ]
  • The RNA THERMO database, version 1.3 [ XML file ] (the T-Full data set containts the data from this XML file)
  • The S-Full-Train data set (including predictions with the Turner99 parameters) [ TXT file ]
  • The S-Full-Test data set (including predictions with the Turner99 parameters) [ TXT file ]
  • The S-Full-Alg-Train data set (including predictions with the Turner99 parameters) [ TXT file ]
  • The S-Full-Alg-Val data set (including predictions with the Turner99 parameters) [ TXT file ]
  • The S-STRAND2 data set (NOT including predictions with the Turner99 parameters) [ TXT file ]
  • The S-STRAND2 data set, but where the pseudoknots are not opened [ TXT file ]
• Code
  • The MultiRNAFold version used in my thesis (that includes Simfold) was 1.12. A slightly older version 1.11 is here [ download page ]
  • The CG version used in my thesis was 2.0. Here's the web page for the older CG 1.0, let me know if you need the 2.0 version (note that in order to run CG you need to have access to the CPLEX licence and one of the CS clusters).
  • The BL version used in my thesis was 2.0. Let me know if you need it.
• Scripts
  • (added on 2009-01-21) To generate the BL* S-Full-Test column in Table 5.8, download this directory, unarchive, cd into it, and type:

    perl test_accuracy.pl icics S-Full-Test/ BLdir/BLstar_params.txt model-363T_all-T.txt BLdir/accuracy.txt

    Note this is using the icics cluster, so you have to be able to start jobs before you use this.

    The average accuracy measures are written in file BLdir/accuracy.txt.