| subset | subset of proteins common to all prediction methods applied | |
| nprot | number of proteins in subset of proteins common to all prediction methods applied | |
| sots | more info | standard-of-truth used to evaluate (i.e. method assigning secondary structure given the three-dimensional structure) |
| rank | Rank of given method. Methods that are not distinguishable since they differ by less than the estimated error given the number of proteins in the common set (ERRsig) are assigned the same rank! | |
| method | more info | Name of prediction method (all methods) |
| Q3 | definition | three-state per-residue accuracy (percentage of correctly predicted residues) |
| ERRsigQ3 |
significant deviation for three-state per-residue accuracy (percentage of correctly predicted residues)
defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| Q3H%o | definition | correctly predicted helix residues (percentage of helix observed) |
| ERRsigQ3H%o |
significant deviation for Q3H%o; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| Q3H%p | definition | correctly predicted helix residues (percentage of helix predicted) |
| ERRsigQ3H%p |
significant deviation for Q3H%p; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| Q3E%o | definition | correctly predicted strand residues (percentage of helix observed) |
| ERRsigQ3E%o |
significant deviation for Q3E%o; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| Q3E%p | definition | correctly predicted strand residues (percentage of helix predicted) |
| ERRsigQ3E%p |
significant deviation for Q3E%p; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| Q3L%o | definition | correctly predicted loop residues (percentage of helix observed) |
| ERRsigQ3L%o |
significant deviation for Q3L%o; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| Q3L%p | definition | correctly predicted loop residues (percentage of helix predicted) |
| ERRsigQ3L%p |
significant deviation for Q3L%p; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| SOV3 | definition | per segment accuracy (three states) |
| ERRsigSOV |
significant deviation for SOV; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| sovH | definition | per segment accuracy for helix residues |
| ERRsigsovH |
significant deviation for sovH; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| sovE | definition | per segment accuracy for strand residues |
| ERRsigsovE |
significant deviation for sovE; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| sovL | definition | per segment accuracy for non-regular residues |
| ERRsigsovL |
significant deviation for sovL; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| BAD | definition | percentage of confused HELIX/STRAND residues (helix predicted as strand, strand predicted as helix) |
| ERRsigBAD |
significant deviation for BAD; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| info | definition |
information index accuracy, with the following convention:
info = (infoO + infoP) / 2 infoO = info as percentage of observed infoP = info as percentage of predicted |
| ERRsiginfo |
significant deviation for info; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| infoO | definition | information index accuracy (percentage of observed) |
| ERRsiginfoO |
significant deviation for infoO; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| infoP | definition | information index accuracy (percentage of predicted) |
| ERRsiginfoP |
significant deviation for infoP; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| corrH | definition | Matthews correlation index for helix |
| ERRsigcorrH |
significant deviation for corrH; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| corrE | definition | Matthews correlation index for strand |
| ERRsigcorrE |
significant deviation for corrE; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| corrL | definition | Matthews correlation index for loop |
| ERRsigcorrL |
significant deviation for corrL; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| class | definition |
correctness of predicting secondary structure class according to DSSP (class = (classO + classP) / 2 classO = info as percentage of observed classP = info as percentage of predicted |
| ERRsigclass |
significant deviation for class; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| classO | definition |
secondary structure class according to DSSP ( |
| ERRsigclassO |
significant deviation for classO; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| classP | definition |
secondary structure class predicted ( |
| ERRsigclassP |
significant deviation for classP; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| contDH | definition |
difference of :
observed content of helix - predicted content of helix |
| ERRsigcontDH |
significant deviation for contDH; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
|
| contDE | definition |
difference of :
observed content of strand - predicted content of strand |
| ERRsigcontDE |
significant deviation for contDE; defined by: ERRsig(x) = SIG(x) / square root (numprot) SIG(x) = standard deviation for variable x |
