Table 2. Glioblastoma patient characteristics.

ProbeID	Fold change	P-value (unadjusted)
hsa-miR-107_st	-0.3050	0.000024
hsa-miR-3125_st	-0.3397	0.000092
hsa-miR-331-3p_st	-1.0500	0.000101
hp_hsa-mir-4315-2_s_st	-0.2556	0.000172
hsa-miR-548x_st	-1.2050	0.000204
hsa-miR-3126-5p_st	-0.6107	0.000227
hp_hsa-mir-885_st	0.3883	0.000292
hsa-miR-4270_st	1.1590	0.000676
hsa-miR-103_st	-0.3822	0.000705
hsa-miR-887_st	-0.7126	0.000873

The ten most significantly deregulated miRNAs in the 38 glioblastomas comparing STS to LTS depicted as fold change relative to LTS. Using the leave-one-out cross validation approach, the model predicted an accuracy of 78% with hsa-miR-107_st, hsa-miR-548x_st, hsa-miR-3125_st and hsa-miR-331-3p_st as optimal predictors (indicated with bold).