[greigite]

Could you give a citation? I'm not familiar with the Turner et al paper you mention.

[kmcarr]

Could we implement a rule along the lines of that if reads start in the same position, but are different in length by more than 1% (could get incorporation errors changing the length of duplicate reads) then they are not duplicates?

This rule would only work if you are sequencing the entire fragment; that is to say the sequence is reaching the 3' adapter and thus you can determine the exact size of the original fragment. This should not be the case. The average fragmentation size should be sufficiently large (500 - 800 bp) such that the 200 cycles of sequencing on the FLX Titanium never reach the 3' adapter. Therefore you would have no way of knowing for a given read what it's underlying fragment size is.

[greigite]

This rule would only work if you are sequencing the entire fragment; that is to say the sequence is reaching the 3' adapter and thus you can determine the exact size of the original fragment. This should not be the case. The average fragmentation size should be sufficiently large (500 - 800 bp) such that the 200 cycles of sequencing on the FLX Titanium never reach the 3' adapter. Therefore you would have no way of knowing for a given read what it's underlying fragment size is.

I think I'm missing something here. If I understood your earlier post correctly (quoted below), shotgun fragments that are randomly sheared at exactly the same location on one end should also have the same sequence length, regardless of the size of the original fragment. Why would it be necessary to know the underlying fragment size?

<snip>
If I have 10 fragments all originating from the 5' end of 10 copies of the same cDNA but all of varying lengths between 500-800 nt, and the 454 adapters are ligated in the same orientation, then I should get exactly the same sequence from all of them. The sequence will start at the 5' end and will stop when the machine has completed its 42 cycles, regardless of how long the inserted fragments are. This is what I mean by "if two reads start.....".

[kmcarr]

Yes, you've got it. Your question (at least as I understood it) was whether or not you could differentiate between independent fragments which happen to share the same 5' end and PCR duplicates by the length of the read obtained. No, you can't. Using 454 cycle sequencing, if two reads start at the same position they will end up being exactly the same length as each other, whether they were originally independent fragments or PCR generated duplicates. This statement relies on an assumption of 454 library preparation that the fragment size if your library is longer than the expected read length. That is the only point I was making about original fragment sizes.

[greigite]

OK, I understand now- very good point and thanks for elaborating. Is there another way you suggest that might work to differentiate between fragments sharing the same 5' end and PCR duplicates? Perhaps there is no way to do it when the shotgun fragments are drawn from a large population of an organism with very little variation, such that even two fragments which happen to share the same 5' end are 100% identical. I guess one has to assume the relative probabilities of PCR duplication versus shearing at the same point, and this probably scales with the coverage (e.g. for lower coverage regions if two reads start at the same point they are more likely to be PCR duplicates).

Yes, you've got it. Your question (at least as I understood it) was whether or not you could differentiate between independent fragments which happen to share the same 5' end and PCR duplicates by the length of the read obtained. No, you can't. Using 454 cycle sequencing, if two reads start at the same position they will end up being exactly the same length as each other, whether they were originally independent fragments or PCR generated duplicates. This statement relies on an assumption of 454 library preparation that the fragment size if your library is longer than the expected read length. That is the only point I was making about original fragment sizes.

[pmiguel]

OK, I understand now- very good point and thanks for elaborating. Is there another way you suggest that might work to differentiate between fragments sharing the same 5' end and PCR duplicates?

If standard shotgun techniques are being used then it should be possible to distinguish between PCR duplication and shearing bias.

If shearing bias is the issue then one expects equal numbers of reads with the shear site against the PA adaptor as against the PB adaptor. This can only be seen in amplicons short enough to be read across into the PB adaptor. But if you are mapping against a reference genome or clustering you would also expect to see an equal number of reads mapping immediately before the shear site (on the opposite strand) as after the "fragile" shear site. If you do not, then you would suspect PCR duplication.

Phillip