A tree-based conservation scoring method for short linear motifs in multiple alignments of protein sequences

BMC Bioinformatics

Table 1 Accuracy of the different models using different sets of parameters

Model	FPR	FNR	D _lim	P _lim
EXC DISC 1	0.22	0.14	0.30	1.00
EXC DISC 2	0.12	0.20	0.50	1.00
EXC CONT 1	0.19	0.17	0.50	0.80
EXC CONT 2	0.26	0.10	0.30	0.80
MIS DISC	0.39	0.16	0.30	1.00

The parameters P_limand D_limdetermine the number of informative sequences considered when calculating the conservation score of each instance (see Conservation Score section for further explanation). In all the cases, the false positive and false negative rates are calculated for the optimal threshold (0.58), which maximises both the model's sensitivity and specificity. FPR = 1 - specificity = $\frac{F P}{F P + T N}$ FNR = 1 - sensitivity = $\frac{F N}{F N + T P}$

ISSN: 1471-2105