Calculate statistics from a set of structures

SOAP statistical table

class statsTable.rawsp(pdbset='', features='', genmethod='', tlib=None, decoy=False, routine='calc_sp_all', model={})[source]

Statistics from a set of structures

Parameters:
  • pdbset (str) – the name of the set of PDB or decoy structures

  • features (str) – the name of the features, see feature

  • genmethod (str) – the filters used for generating the table, see utility.decode_genmethod() for more detail.

  • decoy (str) – True if the rawsp.pdbset represent decoys

  • model (dict) – the parameters for this class can also be passed by a dict

Example genmethod:
  • bs2dsp: bond separation 2 with disulfide bond taken into consideration

  • cs1: chain separation >=1

  • ss1: sequence separation >=1

calc_sp_all(runnumber)[source]

Methods to generate the statistical potential from PDB files in the currnt directory.

class statsTable.scaledsp(pm='', rspo=[], model=[], previouspm='')[source]

Processed statistics table from a set of structures

Parameters:
  • rspo (rawsp) – class rawsp object

  • pm (str) – the method for processing the raw table scaledsp.rspo

  • model (dict) – the parameters for this class can also be passed by a dict

pm definition:

pm  => [pp[0-9]+,+] ,[(gps|ks)\number,+] ,npend|npsum
pp[0-9]+ : add number to raw table
(gps|ks)\number : gaussian process or kernel smoothing with strength \number.
npend|npsum : normalize the count by the last bin or by the sum of the distribution (results in a proper discrete probabilistic distribution).