Illumina adapters bind equally to both strands?

[chipnewb]

When looking at paired-end sequencing:

I noticed that during ChIP-seq, the adapters bind to both ends of a polished double-stranded DNA molecule. If we focus on a particular fragment of DNA, after PCR amplifications, you will have two sets of double-stranded DNA (one of which was originally made from the + strand, and one made from the - strand).

When this get sequenced, you will have two sets of reads. One that was made from the original + strand and one made from the original - strand. Read 1 from the sample obtained originally from the + strand will be the reverse complement of Read 2 from the sample obtained from the - strand, while Read 2 from the sample obtained from the + strand will be the reverse complement of Read 2 from the sample obtained from the - strand.

My question is do programs like BWA take this into account and simply map everything to the + strand? i.e. does BWA take all the reads that map 5'->3' on the - strand, take the reverse complement, and map them to the + strand?

[Chipper]

You are correct, this is a problem that is not adressed by most (all?) peak finders so you may have to increase your cut-off for libraries with low complexity. Bwa will just align the reads. The reason for this duplication of the original fragments is the Y-adaptors used by Illumina.

[chipnewb]

Thank you. I worked out Illumina's process by hand last night and it turns out that for every sheared double-stranded DNA you generate, Illumina will generate a 4 reads (2 for each strand). In the end you generate two reads for each strand, however read1 that is generated for the "top" strand will be the same as read2 for the bottom strand. And read2 that is generated for the top strand will be the same as read1 for the bottom strand.

[chipnewb]

So for example let's say we have the following:

5' GGGGTTTTAAAA 3' (simply for ease of illustration).
3' CCCCAAAATTTT 5'

After Illumina is done with it you will have 4 reads:

read1 = GGGG, read2 = TTTT
read1 = TTTT, read2 = GGGG

Bwa then takes those reads and tries to map them to them to the genome.