TopHat: the results confused me

[Maria_Lu]

Hi, Lifeng,

Thank you so much for you kindness.
I'll calculate these junctions again according to your remind.

[lifeng.tian]

Maria,

Since you did not give the complete junctions.bed line, I attached two lines
from my TopHat analysis:

75nt run
chr20 251862 256677 JUNC00000001 11 - 251862 256677 255,0,0 2 46,69 0,4746


50nt run
chr20 251879 257723 JUNC00000001 2 - 251879 257723 255,0,0 2 29,39 0,5805


Here is the link to BED format at ucsc: http://genome.ucsc.edu/FAQ/FAQformat.html#format1

You can calculate the (start,end) of the junction by:
e.g.for 75nt run, find col 2, 10, 11, 12:
col 2: feature start 251862
col 10: block count 2
col 11: block sizes 46,69
col 12: block starts 0,4746
You can find the splice junction coordinates by:
start: col2+col11.first.item+1
end: col2+col12.second.item,
i.e.,
start: 251862+46+1=251909, end: 251862+4746=256608
chr20:251909-256608


for 50nt run,
junction is : chr20:251909-257684


Now, in this case, TopHat finds two different junctions. You can also
copy and paste the BED directly to ucsc custom track and visualize them.

Hope it helps.

Lifeng

[Maria_Lu]

I used TopHat to find exon-exon junctions.

But the results in the output 'junctions.bed' files confused me.
I seperated RNA-seq data into two datasets as one is 76bp*2 another is 40bp*2, then ran TopHat individually.

However, the two 'junctions.bed' files gave different results.
Here are examples of the two 'junctions.bed' result:
One reports the following junctions:
chromosome12 12302 12721 JUNC00000002 5 -
chromosome12 33389 34997 JUNC00000003 6 +
chromosome12 33688 34964 JUNC00000004 2 +
chromosome12 35474 35675 JUNC00000005 5 +
chromosome12 35718 35949 JUNC00000006 9 +

Another reports the following junctions:
chromosome12 12303 12723 JUNC00000005 26 -
chromosome12 33679 34982 JUNC00000007 3 +
chromosome12 35490 35674 JUNC00000008 6 +
chromosome12 35711 35949 JUNC00000009 7 +

These junction locations of each output file were similar but different.
When I extracted the detailed sequences, no GT-AG was found.

Does anybody know how to explain it?