Can these fusion predictions by tophat be trusted

Question

[Dataminer]

Hi,

I have got these results from tophat-fusion:

chr15-chr17    74317267    38506035    ff    1    13    1    7    43    33    10.000000    @    13 26 38 51 63     @    CCAGCCCAGAGGCTGCCAGCACTCCCAGGGACCCTATTGACGTTGACCTG GTGAGATGGGTTTGAGGTCTGAAGGGGTGGTGGT

GGGGCCCAGCAGGTCA    @    TGCTCCCACTGTGGGTGTGGACAACCTGACTCCCTCCCCTCCATACCCAG GGCTTCTTCCGCCGCAGCATCCAGAAGAACATGGTGTACACGTGTCACCG    @    1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0
 0 0 0     @    1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0     @    -8418:28 -8481:-1326 -8394:-1415 -8411:-1464 -8421:-1464 -8479:-1417 -8461:-1462 -8476:-1471 -8794:-5041 -872
9:-5209 -8803:-5220 -8834:-5250 -8799:-5285
chr15-chr17    74325754    38504567    ff    5    13    9    1    54    39    15.480002    @    14 30 41 53 68     @    GGTCTTCCTGCCCAACAGCAACCACGTGGCCAGTGGCGCCGGGGAGGCAG GTAGGGAGGTGGGTAGGGCAGTGGCCTGGGTGCT
GCCCGCCAGGGTCTGG    @    AAGGATGGCGACCTAGGTCTCTAACTGCCCCTCCCCTCTTCTCTCTCTAG CCATTGAGACCCAGAGCAGCAGTTCTGAAGAGATAGTGCCCAGCCCTCCC    @    5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 4
 4 4 3     @    5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 3 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0     @    11:-3 26:6 8:51 66:4 76:4 93:53 6:142 69:1496 -307:-3573 -242:-3741 -316:-3752 -347:-3782 -312:-3817
chr15-chr17    74326139    38500724    ff    8    13    10    3    63    60    1.859375    @    15 27 41 55 67     @    GCATACCATAACAGGAAACTGGACTCCCCATGACCTTAACTCTGCTCTCT CAGTCTACACCCATCATTTCCCCAAGTGTTTCTG
CAAAGGCCACCTACCT    @    GGGCTGCCAGAGTCACCCCTTCCACTGCCTTGGCCACCTTCTCCAGAGGG CTGGAGAGAAGCTGGGATCTGAGACCTTGGTCTCCAGCCCCTGTCTCTTC    @    8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 7 7 7 7 7 6 6 6 6 6 6 6 6 6 6 6 6 5 5 5 5 4 4 4 4 4 4 4 4 4 4
 4 4 4     @    8 8 8 8 8 8 8 8 8 8 8 8 8 5 5 5 5 5 5 5 5 5 5 5 5 4 4 4 4 4 4 4 4 4 4 4 4 4 4 3 3 3 3 2 2 2 2 2 2 2     @    38:61 73:26 69:91 143:102 78:270 396:3840 411:3849 393:3894 451:3847 461:3847 478:3896 391:3985 454:5339
chr15-chr17    74326818    38487647    rr    15    4    18    1    8515    61    0.675556    @    13 21 35 46 57     @    GTTTTCGGCATCTGAGTCTTCCGAGCTGCTGATCACCACAACGCGTTCCT CTGAAACGGGGAAGGGGAGATGCATTATGAACTG
CCTAGGATGCATGCTG    @    TGCCACCCTCCACAGTCCCTTCCCACCCCAGCCCCCACACCCAGACTTAC TGGCTGGGGATGGTGTGCTATATCCACTAACTGGAAGCTGGTGCTGGAGA    @    15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 12 12 12 12 12 12 12 10 10 10 10 10 10 10 9 9 9
9 8 8 6 6 6 6 6 6 6 6 6 5 5 5 5 5 4     @    15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 14 14 14 14 13 13 12 11 10 10 10 10 10 9 9 9 9 9 9 9 9 9 7 7 6 6 6 6 5 5 5     @    23:8 44:21 8638:4 8622:44

chr15-chr17    74325591    38504567    ff    2    1    2    26    34    47    10.500000    @    15 28 44 56 66     @    CCCAGTGCCCCAGGAAGGTCATCAAGATGGAGTCTGAGGAGGGGAAGGAG GCAAGGTTGGCTCGGAGCTCCCCGGAGCAGCCCA
GGCCCAGCACCTCCAA    @    AAGGATGGCGACCTAGGTCTCTAACTGCCCCTCCCCTCTTCTCTCTCTAG CCATTGAGACCCAGAGCAGCAGTTCTGAAGAGATAGTGCCCAGCCCTCCC    @    2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
 0 0 0     @    2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 0 0 0     @    -106:4

I have the feeling that the number of reads spanning the fusion is too low and predictions might not be accurate, but, on the other hand we were expecting these fusions but not with low number of reads. More over the the number of contradicting reads are also high.

I am really confused, what shall I do? can I take these results and confirm that fusion takes place or it is just by chance that I am seeing this.

Please help with your expert opinion and how you would have judged this result.

Thank you

[Manvendra Singh]

I have a feeling from my analysis that TopHat fusion gives so many fusions and lot of false positive results, whereas STAR is better , you can even have trust on MapSplice2.

Answer 1

[Manvendra Singh]

I would suggest you to use STAR and mapsplice2 (it would take little time) and take overlapping fusions into the account, it would be trusted

Answer 2

[bstrs]

Have you tried FusionCatcher? FusionCatcher has smallest amount of false positive fusions (see: https://code.google.com/p/fusioncatcher/wiki/comparison)

[Manvendra Singh]

Yes, but they have not compared STAR and mapsplice2,

I personally trust more on these two tools for chimeric transcript detections (gene-fusions)

[enxxx23]

Actually STAR and mapsplice2 perform very poorly very for gene finding in RNA-seq. I trust FusionCatcher and SOAPFuse the most. It is very very easy to finding fusion genes in RNA-seq and have an exceptionally good sensitivity. The challenge is to have a very good specificity while still finding fusion genes in RNA-seq and this is were FusionCatcher is really good.

See (in STAR google forum search for "STAR-Fusion for finding fusion transcripts in cancer transcriptomes"):