hello!

I using package rtracklayer to processing some paired end reads.

using export function, I can convert my data to bed format, but there is a question confused me.

I using bedtools bamtobed function to convert bam file to bed file, the output are as follows:

chr10 29334602 29334652 SRR891270.293/2 255 +

chr10 29334606 29334656 SRR891270.293/1 255 -

chr10 61751512 61751562 SRR891270.256/2 255 +

chr10 61751525 61751575 SRR891270.256/1 255 -

chr10 103211892 103211942 SRR891270.153/2 255 +

chr10 103212015 103212065 SRR891270.153/1 255 -

then , I want using MACS2 to call peaks, so I have to merge these reads and the output I think should be like this:

chr10 29334602 29334656 SRR891270.293 255

chr10 61751512 61751575 SRR891270.256 255

chr10 103211892 103212065 SRR891270.153 255

no strand infomation exist because of merge.

but when I using export function of rtracklayer, the output are

chr10 29334602 29334656 SRR891270.293 0 -

chr10 61751512 61751575 SRR891270.256 0 -

chr10 103211892 103212065 SRR891270.153 0 -

the 5th cloumn is not important, but why the strand infomation still exist?

from what I learned, the paired end principle  is like 

---------------------------[400bp]-----------------------------------

--['+':50bp]-->                                            <--['-':50bp]--

when you merge the 2 reads, it means you get the total reads from its start(the small number) to its end(the large number), why strand info still exist?

Is the package "rtracklayer" wrong or I miss some thing during the merge?    

thanks!