Systematic detection of transcript usage from Whippet results

Question

[graeme.thorn]

Is there a systematic (and programmatic) way of determining which transcripts are differentially expressed given a set of Whippet events for this gene?

For instance, if the output of a gene is the following:

ENSG00000116266 1   chr1:108746674-108746681    +   TS  0.57664 0.12668 0.44996 0.926   K1  0.9999
ENSG00000116266 2   chr1:108746682-108746786    +   TS  0.42267 0.87175 -0.44908    0.907   K1  0.9999
ENSG00000116266 3   chr1:108752257-108752306    +   CE  0.99711 0.99793 -0.00081905 0.575   K0  0.0
ENSG00000116266 4   chr1:108753063-108753144    +   CE  0.99753 0.99843 -0.00090097 0.632   K0  0.0
ENSG00000116266 6   chr1:108756690-108756766    +   CE  0.99739 0.99845 -0.001059   0.611   K0  0.0
ENSG00000116266 8   chr1:108758510-108758588    +   CE  0.99629 0.99831 -0.0020177  0.604   K0  0.0
ENSG00000116266 9   chr1:108759985-108760085    +   CE  0.99758 0.99862 -0.0010393  0.603   K0  0.0
ENSG00000116266 10  chr1:108772665-108772819    +   CE  0.99822 0.99874 -0.00051795 0.571   K0  0.0
ENSG00000116266 11  chr1:108776333-108776423    +   CE  0.99815 0.99853 -0.00037609 0.548   K0  0.0
ENSG00000116266 12  chr1:108779286-108779410    +   CE  0.99794 0.99816 -0.00022023 0.543   K0  0.0
ENSG00000116266 14  chr1:108782422-108782517    +   CE  0.99521 0.9984  -0.0031897  0.602   K2  0.3733
ENSG00000116266 15  chr1:108782649-108782706    +   CE  0.99816 0.9984  -0.00023314 0.518   K0  0.0
ENSG00000116266 18  chr1:108794827-108794907    +   CE  0.99783 0.99803 -0.00020365 0.536   K0  0.0
ENSG00000116266 19  chr1:108796234-108796372    +   CE  0.99797 0.99832 -0.00035294 0.507   K0  0.0
ENSG00000116266 20  chr1:108796620-108796726    +   CE  0.99825 0.99864 -0.00039527 0.543   K0  0.0
ENSG00000116266 21  chr1:108798145-108798237    +   CE  0.99819 0.99843 -0.00024271 0.506   K0  0.0
ENSG00000116266 22  chr1:108800220-108800305    +   CE  0.99822 0.99842 -0.00019266 0.503   K0  0.0
ENSG00000116266 23  chr1:108807401-108807549    +   CE  0.99    0.99869 -0.0086972  0.768   K1  0.2364

where there is an obvious event of switching transcription start sites, is there a way of systematically determining which are the most likely transcripts used in condition A and condition B?