Reads of Insert - PacBio, Forward or Reverse?

[Fad2012]

Dear all,

I am wondering when the analysis software like smrtPORTAL generate the Reads of Insert from the subreads, in which orientation those reads would be ( I mean the reads of Insert reads). And if some of them in reverse, is it reverse or reverse complement. And is there any way to find out how many reads of each orientation you have?

I am asking because, in my data around 14000 Reads of Insert, I have a significant number of the reverse reads, i can see them appearing on the IGV when i got the sam file. And want to know if whether each Read of Insert represents one molecule, or I have to count the Reverse as well, so will have 2 reads for one molecule. In this way of calculating it, then i would have only 7000 reads that represent 7000 individual molecules???

Many thanks

[Brian Bushnell]

Around half should be "forward" and half should be "reverse". Each molecule should generate one read of insert, not two.

[Fad2012]

Reference

Dear Brian, thank you for your answer.

Are you aware about any possible reference where I can find more detailed information about the process of producing forward and reverse reads, and how this can be treated in the context of the downstream analysis.

Many thanks

[Brian Bushnell]

All normal high-throughput DNA sequencing produces 50% of reads forward and 50% reverse; so, downstream analysis tools are designed to cope with that, typically by treating the strand as unimportant. There are some notable exceptions, such as strand-specific RNA-seq.

In some cases, like when calling variations after alignment, requiring the variation to have been seen both on a plus and minus mapped read can be used as a heuristic to reduce false positives, but for the most part, the strand is ignored.

PacBio is a little different because the molecules can be read multiple times, alternating between plus and minus. Probably, the strand of the read of insert is the strand of the first complete read-through, but it doesn't really matter, and of course the software has no way of knowing which strand is which since it depends on the reference.

As for references for reading further... hmmm, maybe someone else will suggest one?

[rhall]

For a single molecule (SMRT Bell) sequencing can start anywhere, therefore for a single consensus (read of insert) the chances of being forward or reverse are 50/50. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2926623/

[Fad2012]

Thank you very much