Table 2. Fusion transcripts detected by the RACE-seq approach.

Gene A	Gene B	Break position A	Break position B	Split reads*	Spanning reads*	In Frame	Probability score†	Validated§	Sample
VTI1A	TCF7L2	Chr10: 114,220,341	Chr10: 114,900,943	1234	163	Yes	0.96	Previously reported	NCI-H508
TCF7L2	RP11–57H14.3	Chr10: 114,799,885	Chr10: 114,648,494	114	41	No	0.95	Previously reported	HCT-116
VWA2	TCF7L2	Chr10: 116,014,807	Chr10: 114,900,943	206	67	Yes	0.98	Yes	Sample 17_T
DHX35	BPIFA2	Chr20: 37,632,550	Chr20: 31,767,410	862	4	Yes	0.23	Yes	Sample 16_T
CASZ1	MASP2	Chr1: 10,820,757	Chr1: 11,090,938	107	NA	Yes	NA	Yes	Sample 4_T

Both previously known as well as novel fusion transcripts were detected by the RACE-seq approach and subsequent analysis of paired-end sequencing reads. Only the top-scoring nominated candidates are shown for each fusion pair. The known as well as the novel fusion transcripts VWA2-TCF7L2 and DHX35-BPIFA2 were nominated by the fusion detection software, whereas the novel CASZ1-MASP2 was detected by the Unix utility grep and scaffold realignment.

^*The number of split reads and spanning reads supporting the fusion junction, determined from the fusion detection software or scaffold realignment for the CASZ1-MASP2 fusion (split reads contain the fusion boundary in the read itself, whereas spanning reads are paired ends that harbor the fusion boundary within the insert sequence).

^†Probability score determined by the fusion detection software.

^§Two fusion genes involving TCF7L2 have previously been reported and served as positive control fusion targets for the RACE-seq. The novel fusions were validated by Sanger sequencing.