. 2009 Aug 28;10:271. doi: 10.1186/1471-2105-10-271

Table 3.

Models produced by Stepwise Binary Logistic Regression Iteration and M2 Cross-Validation.

`SBLR Model`	`M0`	`M1`	`M2`	`M3`	`M2 Cross-Validation`
`Training Set` `Deletion`	`none`	`CT665` `CT681a` `CT681b` `CT743`	`CT665` `CT681a` `CT681b`	`CT681a` `CT681b`	`CT665` `CT681a` `CT681b`

`Variables in Model^a`	`+HMM_SCORE` `+STABLE1_37` `-POSITION` `+CURVE_L32` `-GC_L128` `+RIGID_L96` `+CURVE`	`+HMM_SCORE` `+STABLE1_37` `-GC_L32` `-POSITION` `+CURVE_L32` `-CURVE_L64` `-GC_L128` `+TWIST`	`+HMM_SCORE` `+STABLE1_37` `-POSITION` `+CURVE_L32` `-STABLE_L32` `+SIDD_H24` `-CURVE_L128` `-SIDD_L128` `+RIGID_L96`	`+HMM_SCORE` `+STABLE1_37` `-POSITION` `+CURVE_L32` `-STABLE_L32` `-STABLE27_37` `+CURVE`

`Sensitivity or Recall`	`19/29` `(0.655)`	`25/25` `(1.0)`	`26/26` `(1.0)`	`25/27` `(0.926)`	`23/26` `(0.885)`

`Specificity`	`2426/2428` `(0.999)`	`2083/2083` `(1.0)`	`2226/2226` `(1.0)`	`2322/2323` `(1.0)`	`2215/2226` `(0.995)`

`PPV or` `Precision`	`19/21` `(0.905)`	`25/25` `(1.0)`	`26/26` `(1.0)`	`25/26` `(0.962)`	`23/34` `(0.676)`

`NPV`	`2426/2436` `(.996)`	`2083/2083` `(1.0)`	`2226/2226` `(1.0)`	`2322/2324` `(0.999)`	`2215/2218` `(0.999)`

`Accuracy`	`2445/2457` `(0.995)`	`2108/2108` `(1.0)`	`2252/2252` `(1.0)`	`2347/2350` `(0.999)`	`2238/2252` `(0.994)`

`AUC^b`	`0.995`	`1.0`	`1.0`	`0.999`	`0.992`

^aThe variables are listed in order of entrance into the model and the sign indicates the sign of the coefficient.

^bROC analysis Area Under the Curve