Maximum fragment size

[matth431]

Last thread on a similar topic was back in 2008 so figured I'd check if anything has changed - to save time whilst performing a genomic library prep (TruSeq) I decided to try a 0.6x AMPure cleanup instead of gel-cutting for size-selection. Resulting libraries are a bit bigger than I was expecting - around 700bp enriched product so insert size of ~600bp! Are fragments this size going to cluster correctly?

I suppose I could burn a MiSeq run to find out before risking them on the HiScan... but any advice appreciated!

[TonyBrooks]

They should still cluster OK. If you want to err on the side of caution, you can load at a slightly lower density to compensate for the marginally larger clusters. Although honestly, I don't think that they are sufficiently larger than normal to cause a problem. This only really seems to happen when you get over 1kb inserts.
Remember than Illumina will (probably) be rolling out 400bp PE on MiSeq soon which will need >700bp inserts to get the most out of the sequencing.

[matth431]

They should still cluster OK. If you want to err on the side of caution, you can load at a slightly lower density to compensate for the marginally larger clusters. Although honestly, I don't think that they are sufficiently larger than normal to cause a problem. This only really seems to happen when you get over 1kb inserts.
Remember than Illumina will (probably) be rolling out 400bp PE on MiSeq soon which will need >700bp inserts to get the most out of the sequencing.

Many thanks - I did wonder regarding the 2x400bp reads but assumed users might also want a bit of overlap between the two reads for error-checking. I have a freebie MiSeq kit I can use to check, will perhaps try a 7pM load to space the clusters a bit.

[pmiguel]

For a more extreme example see this thread.

--
Phillip

[matth431]

Thanks again - looks like it should work then but possibly exclude some of the larger inserts... that's not really an issue for these samples (just doing some genomic shotgun to look for SNPs) but worth bearing in mind for future.

Interesting that the 0.5x AMPure in that case gave such huge fragment sizes - wouldn't have expected such a difference (nearly 3-fold bigger!) to the range I got here, although my DNA was sheared for 500bp mean size...

[hawaii454-0]

Hello,
What do you mean by a 0.6x ampure clean up? 1:0.6 ratio?

Thanks,
Anna.

[TonyBrooks]

If your reaction size is 100uL, then add 60uL of Ampure beads during clean up as opposed to the standard 180uL (1.8x). Keep the ratio the same (30uL of beads for 50uL reaction clean ups).
By reducing the volume of beads, you reduce the buffer concentration which moves the binding titration point up (lower salt conc in the reaction means only more negatively charged DNA (i.e. longer) bind). The beads themselves shouldn't be an issue as 1µL of beads could potentially bind >1ug of DNA.

[hawaii454-0]

Thanks Tony.

Does anyone have final libraries which show shoulder peaks on their larger insert libraries?

Thanks,
Anna.

[pmiguel]

If your reaction size is 100uL, then add 60uL of Ampure beads during clean up as opposed to the standard 180uL (1.8x). Keep the ratio the same (30uL of beads for 50uL reaction clean ups).
By reducing the volume of beads, you reduce the buffer concentration which moves the binding titration point up (lower salt conc in the reaction means only more negatively charged DNA (i.e. longer) bind).

The salient component of the "buffer" here is PEG (with NaCl also at high concentration--polynucleotides don't seem to precipitate as free acids, but do as salts). Like alcohols, PEG is said to "occupy" some percentage of the solvation capability of water. The effect is that longer molecules come out of solution preferentially. As more and more of the solvation capacity of water is devoted to PEG (or alcohol), shorter and shorter molecules of polynucleotides can be forced from solution.

The binding of this precipitant to the carboxyl groups on the surface of these magnetic beads is another issue, and one that I have not seen an explanation for. But it does seem that this surface has some affinity for polynucleotides. But really it is just a faster method of collecting your precipitate than a hard spin in a centrifuge.

I have seen at least one report on seqanswers of using an alcohol precipitation with an Ampure beads collection of the precipitated DNA.

--
Phillip

[captainentropy]

The binding of this precipitant to the carboxyl groups on the surface of these magnetic beads is another issue, and one that I have not seen an explanation for.

Once I read this post on core-genomics it began to make sense to me how it's working

http://core-genomics.blogspot.com/2012/04/how-do-spri-beads-work.html?showComment=1372864042066#c8174070126511809493

[pmiguel]

Once I read this post on core-genomics it began to make sense to me how it's working

http://core-genomics.blogspot.com/20...70126511809493

Yes, later I quoted the same comment in another thread, here. At the time, I was less than convinced by the commenter's argument. However, re-reading the comment now, I am liking it more.

The original SPRI paper, Nucl. Acids Res. (1994) 22 (21): 4543-4544 doi:10.1093/nar/22.21.4543, made it sound like the carboxyl surface of the beads being used was what bound the DNA. But if the authors merely tried a few surface derivatives of these beads and went with the one that gave the highest yields then there is another explanation. That is that any beads derivatized with positively charged molecules would more-or-less irreversibly bind the DNA and therefore appear to give a poor yield. The carboxyl groups would weakly repel the DNA and hence allow for easy elution.

This is basically what the anonymous commenter you link to is writing.

--
Phillip