Illumina Paired End reads are complementary

[mchandra]

hi, i'm new to HTS sequencing and I had a couple questions about some illumina data.
I was given two files (_R1 and _R2), which i assumed were each one half of the paired reads. I imported them into CLC genomics where the paired ends were paired. My boss is not convinced that the 1 and 2 designations are the paired ends of the sequences, because the 2 read is complementary to the first. I hope i could get some help on figuring out the paired end reads and how to analyze the results, because the samples were sonicated before the run and 100base pair reads that are complementary doesn't seem to fit...

[GenoMax]

R1 and R2 are file designations of the two reads from each end of the fragment(s).

What do you mean by they are complementary? R1 and R2 pair for a fragment can only be expected to overlap (in the middle) if the insert size was appropriate compared to the cycles (e.g. 300 x 300 bp with 450 bp inserts) used for the run. Otherwise they would be sampling the two ends of a fragment, without any overlap.

[mchandra]

thanks for responding, my impression is that they would not overlap as well. Until i found a couple results in which both 100 bases of R1 and 100bases of R2 were complementary to each other. I had thought it would be because the fragment size was smaller than 100base pairs but because the sample was sonicated i did not believe this could be correct. I wonder though, since I used CLC genomics to upload it and match the paired ends-could i have been wrong in doing that? The read identifiers are listed as ..._1 and ..._2 so I believed they were paired

[GenoMax]

If you did have 100 bp inserts (or there abouts) then they could potentially overlap. Since the size selection is not perfect you may get some fragments that are shorter then you expect.

Based on your description that looks like paired-end data. If you are aligning to a reference then you can verify that the distance between _1 and _2 is consistent with the insert size you expect.

CLC should not overlap the reads unless they really are complementary. There are tools to collapse paired-end overlapping reads into a long representation of the two reads (when you really expect them to overlap). Search for FLASH/BBMerge on the forum to find those.