Skip to main content
. 2015 Apr 17;16:122. doi: 10.1186/s12859-015-0557-5

Table 2.

F-measure [in %] for spliced reads with different number of spanned junctions for simulation 1

Program Number of junctions spanned
1 (13808336) 2 (598297) 3 (11781) 2* (548382) 3* (6908)
CM Bwt1 91.47 14.24 - 15.16 -
CM Bwt2 94.03 78.47 50.21 82.37 72.66
CM Bwa 95.03 82.73 53.33 86.67 76.46
CM Bwa ann 95.74 84.65 53.9 88.47 76.79
MapSplice 2 92.42 79.18 27.27 80.65 3.44
STAR 1-pass 77.63 30.91 5.01 31.91 1.49
STAR 1-pass ann 93.55 81.65 75.71 82.57 82.17
STAR 2-pass 95.0 85.55 82.07 86.59 87.29
STAR 2-pass ann 95.07 86.28 82.55 87.02 86.49
TopHat1 87.83 77.51 63.57 80.42 75.56
TopHat1 ann 88.02 78.99 68.13 81.06 76.1
TopHat2 91.71 87.0 76.92 89.66 88.51
TopHat2 ann 94.84 90.79 85.92 92.05 90.35
GSNAP 83.13 43.45 18.52 42.35 12.29
GSNAP ann 96.47 88.59 79.51 89.86 84.67

*The last two columns show results only for reads for which all exons except the first and last exon had length ≥ 20 nt. For this evaluation, read alignments were only considered a true positive if all simulated splice junctions in the read were recovered and no additional splice junctions were identified. Indels were ignored for this purpose. Recall and precision values for both simulations can be found in Additional file 1: Table S5.