Figure 5

Receiver operator curve analysis of candidate biomarkers miR 21, miR-382 and miR-885 (panels A - C respectively). Data are normalised expression values generated by duplicate miRNA PCR reactions, controlled for reverse transcription efficiency and normalised to miR-191. Animals were grouped into those without any histological evidence of FD (n = 16) and those with histological evidence of FD at any grade (n = 13). MicroRNA-21 (miR-21) was predictive of FD incidence with a sensitivity of 92.31% and specificity of 93.75% (AUC = 0.97). MicroRNAs-382 (miR-382) and miR-885 revealed similar predictive abilities with sensitivity and specificity scores of 69.23% and 93.75% respectively (AUC = 0.89 and 0.86 respectively). Sensitivity and specificity are displayed graphically by the green line. The orange line denotes the line of identity which highlights the shape of the curve should the biomarker offer no discrimination between control and cases. Panels D - F display corresponding expression values for each individual animal ranked from the lowest to the highest expression value. Incidence and severity of FD in the cervical subcutaneous tissue is denoted by the colour of each data point; Grey Circles - no evidence of FD, Green Circles - mild FD; Orange Circles - moderate FD; Red Circles - severe FD. The cut-off value at which maximum sensitivity and specificity is obtained for each biomarker is marked by a red bar (above threshold for miRs 21 and 382, and below threshold for miR-885).