Tophat junctions.bed 1-3bp fluctuate boundary

[pengchy]

Hi,

I have run with tophat2 for many samples seperately and got junctions.bed files. When I combined all the junctions.bed file together to get the information of junction boundary and coverage score, I found the boundary between samples fluctuate with 1-3bp very often. Here are examples:

C189205324      196     484     JUNC0001692     0       -
C189205324      199     770     JUNC0001693     0       -
C189205324      203     775     JUNC0001694     0       -
C189205324      206     770     JUNC0001695     0       -
C189205324      210     771     JUNC0001696     0       -
C189205324      213     772     JUNC0001697     0       -
C189205324      219     774     JUNC0001698     0       -
C189205324      220     775     JUNC0001699     0       -
C189205324      220     804     JUNC0001700     0       -
C189205324      470     787     JUNC0001701     0       -
C189205324      470     817     JUNC0001702     0       -
C189205324      470     818     JUNC0001703     0       -
C189205324      470     821     JUNC0001704     0       -
C189205324      470     823     JUNC0001705     0       -
C189205324      470     824     JUNC0001706     0       -
C189205324      470     827     JUNC0001707     0       -
C189205324      470     830     JUNC0001708     0       -
C189205324      474     784     JUNC0001709     0       -
C189205324      478     782     JUNC0001710     0       -
C189205324      490     777     JUNC0001711     0       -
C189205324      490     779     JUNC0001712     0       -
C189205324      490     782     JUNC0001713     0       -
C189205324      490     787     JUNC0001714     0       -
C189205324      490     798     JUNC0001715     0       -
C189205324      492     774     JUNC0001716     0       -
C189205324      492     775     JUNC0001717     0       -
C189205324      496     775     JUNC0001718     0       -
C189205324      497     775     JUNC0001719     0       -
C189205324      499     795     JUNC0001720     0       -
C189205324      500     785     JUNC0001721     0       -
C189205324      501     797     JUNC0001722     0       -
C189205324      851     1028    JUNC0001723     0       -

The header of these columns are: scaffoldID start end junctionID score strand. the scores were set to 0.
I checked manually for different samples found that different samples have different boundary.

[pengchy]

C183745259      41      302     JUNC00000001    76      -       41      302     255,0,0 2       89,89   0,172
C184664061      52      363     JUNC00000002    404     +       52      363     255,0,0 2       89,89   0,222
C184770029      92      344     JUNC00000003    2567    +       92      344     255,0,0 2       89,89   0,163
C184908199      10      217     JUNC00000004    9       -       10      217     255,0,0 2       54,45   0,162
C185216318      146     332     JUNC00000005    2       +       146     332     255,0,0 2       49,60   0,126
C185216318      130     378     JUNC00000006    119     +       130     378     255,0,0 2       87,89   0,159

I have checked the bed again and found that the showed position in the igv shift according to the column 11. Sorry for the missunderstand.