Keywords for writing automatic parsers for PredictProtein output

Most of the PredictProtein output is designed to be human readable. Consequently, keywords used to automatically parse the output may change. In order to maintain an automatic parser for the PHD predictions, I suggest the following compromise:

Parsing PHD predictions

Sacrifice details of PHD predictions
- Use the output option 'return column format' to obtain the prediction additionally in a column based format (one row per residue).
- Keywords used for the column format returned will remain unchanged:
  AA,PSEC,PACC,RI_SEC,RI_ACC
Keep details of PHD predictions
- Use the output option 'return graph' to obtain the prediction additionally in a column based format (one row per residue).
- Keywords used for the column format returned will remain unchanged:
  No,AA,PHEL,RI_S,OtH,OtE,OtL,PACC,PREL,RI_A,Pbie,Ot0,Ot1,..,Ot9

Parsing alignments

Use the output option 'return msf' to obtain the alignment in the MSF format.

Regular expressions used before sections of output

Sequence alignment header:


The alignment that has been used as input to the network is:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

________________________________________________________________________________

Sequence alignment in MSF format:


--- MAXHOM ALIGNMENT: IN MSF FORMAT
MSF of:

Results from threading alignment (TOPITS):

The HSSP formatted alignment of remote homologues:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

________________________________________________________________________________

PHD prediction header:



The resulting network (PHD) prediction is:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

PHD prediction header:


protein:       t1_1169        length      253
protein:       xxxxxxx        length    nnnnn

note: xxxxxxx name, nnnnn number

Results from EvalSec:

The resulting scores for accuracy are:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

________________________________________________________________________________