TopHat2 - impact if insert size?

[vbernard]

Hello
We analyse Solid 5500 RNA seq reads. I used tophat2 for the alignment. We performed first the mapping with different options for the insert size and std deviation
-r/--mate-inner-dist <int> This is the expected (mean) inner distance between mate pairs. For, example, for paired end runs with fragments selected at 300bp, where each end is 50bp, you should set -r to be 200. The default is 50bp.
--mate-std-dev <int> The standard deviation for the distribution on inner distances between mate pairs. The default is 20bp.

Whatever the size / std dev was, we obtained exactly the same results, meaning these options seems not to be considered.
Please let me know if someone else already saw that and how to fix it. Thanks a lot for your time

[vbernard]

hello
We do not understand why most of the parameters have absolutly NO impact (same results for -T for example)
Anybody else saw that pleass?
Thanks

[Mchicken]

Hey everybody,

has this problem been solved?

I am running TopHat2 version 2.0.13 and use default options with --mate-inner-dist 50 and --mate-std-dev 20.

When I draw 2x 100bp with a distance of 100kbp from my reference genome and map them with TopHat2 as paired end reads I get the following:

Reads are aligned with bit flags 97 and 145 meaning
97 Read is paired -- Mate mapped reverse -- Read is first in pair
145 Read is paired -- Read mapped reverse -- Read is second in pair

and

Left reads:
Input : 1
Mapped : 1 (100.0% of input)
Right reads:
Input : 1
Mapped : 1 (100.0% of input)
100.0% overall read mapping rate.

Aligned pairs: 1
100.0% concordant pair alignment rate.


Why are is the read pair reported as concordantly aligned although I used the default --mate-inner-dist 50 and --mate-std-dev 20 option?

When I use a distance of 1mbp the bitflags stay the same but the reads are at least reported as discordantly aligned.



What I also observed when I align the first 100 read pairs of some real data I saw the following:

Aligned pairs: 9
of these: 1 (11.1%) have multiple alignments
9 (100.0%) are discordant alignments
0.0% concordant pair alignment rate


But when I check the sam-file i see only 8 read-identifiers appearing two times. One read was mapped with bit flag 177 (Read is paired -- Read mapped reverse -- Mate mapped reverse -- Read is second in pair) which means that the mate was mapped reverse. But this mate is not present in the sam-file? Does anyone has a clue?

[Brian Bushnell]

I assume that Tophat thinks there might be an intron in between the reads... perhaps if you adjust the max intron length parameter as well, it will change the results.

[Mchicken]

Thanks for your response Brian. This was indeed the reason for the concordant alignment with 100kbp distance.

But I still have the problem with the 177 bitflag, which tells me that the mate was mapped reverse but is missing in the SAM File.

And what I am also not aware of is how can I see from the SAM file if a read pair aligned concordantly or not? I always get the same SAM file reports regardless of whether I take 100bp long read pairs in a distance of 1kb or 10mbp, where the latter one should not be a concordant alignment.


Cheers,
Mario

[Mchicken]

To the last issue:

According to http://broadinstitute.github.io/pica...ain-flags.html a hook at the second position ("read mapped in proper pair") is what I was looking for. And here also the stuff with "--mate-inner-dist 50 and --mate-std-dev 20" seems to work

[Brian Bushnell]

The flag '177' should not be present if the mate is missing from the sam file; sounds like a bug. Bit 0x2 essentially means "proper pair" while 0x8 means "mate unmapped". 0x8 is unset so the mate should be present and mapped.