need help with library prep

[seqgirl123]

can someone answer my doubts because I would like to improve my protocol:

i shear 5ug of genomic dna on the covaris to 400bp. i get a good fragment size distribution in the 400bp-500bp range, where most of the material is concentrated in this range. there is some gDNA in the higher MW range and low MW range. however, i hope that the adapter-ligation sizeselection step later on eliminates that. using qiaquick after this fragmentation step cleans up some of this low MW dna.
next, i perform end repair, purify with qiaquick column and run some eluted end-repaired dna on a gel. there looks to be plenty here. next, i do 3'dA extension, purify with minelute column, and run some eluted 3'dA extended material on a gel.

questions here are:
-though it may not be discernable on a gel, there will be a shift in bp on the
3'dA gel compared to the end-repaired gel correct, due to the one base addition of A? so the 3' dA gel has a slight shift upwards? does anyone know exactly by how much?

next, i perform adapter ligation and purify with qiaquick column. i run some of the purified material on a gel for sizeselection. i see most of the material is concentrated in the 400-500bp region, so this is good. i do see some higher and lower MW material.

questions here are:

the lower MW material, if below 200bp is what i can safely assume is adapter dimers, single adapters that didn't ligate, ssDNA, or unligated unpolished dna that did not correctly get cleaned up in earlier steps correct? is that all? and what is the higher MW material? it could be unligated unpolished dna that did not get cleaned up, but could it be anything else? would size-selecting after fragmentation reduce this issue, or would the final yield of library reduce if size-selection after fragmention was first performed (instead of at the adapter-ligation step)?

another question is about chimeric fragments. i've read sizeselection after fragmentation and also after adapter ligation reduce chimeric fragments, which is two sizeselections in one prep, so how good would the final library yield be? has anyone optimized their genomic library prep protocol to reduce chimeric fragments, and how so?

fianlly i sizeselect the 400-500bp by cutting out a gel slice, purify the gel slice, and do 18 cycle PCR. i see a band at the 400-500bp range, so the PCR worked.

can anyone address these questions? they would be of great help to me. please don't just refer me to the literature papers such as those from Quail et. al in Nature Methods. I've read them, and none address the questions I am asking here.

[GW_OK]

1. Technically there will be a shift upwards in after A tailing, but unless you're running a gel with 1bp resolution you shouldn't notice a shift at all. Especially a normal agarose gel. Heck, even the bioanalyzer can be plus or minus 5-10bp.
2. Adapter dimers should be less than 100bp in length, given that the adapters themselves are ~33bp in length. These should be mostly taken care of by the minelute columns, which I believe have a size exclusion limit of ~75bp.
3. After ligation you should see your band shift upwards ~50bp. If you don't see a band shift you might not have properly ligated adapters.
4. For size selection yields the Qiagen columns run ~80% yield each, when everything works right. Agencourts SPRI beads are typically higher, ~90+%, but take a lot longer to run. Illumina has some new library tech coming down the pipe which should remove several purification steps to help increase overall yield.
5. 18 cycles for PCR seems a bit high. Are you worried about PCR artifacts? You could probably get away with 10 or less. You only need ~2ul of a 10nM library in the end.
6. Going back to your shearing, are you really seeing genomic DNA after Covaris shearing?
7. Do you have access to a bioanalyzer? They're really helpful for this sort of thing.

[pmiguel]

Originally Posted by GW_OK

1. Technically there will be a shift upwards in after A tailing, but unless you're running a gel with 1bp resolution you shouldn't notice a shift at all. Especially a normal agarose gel. Heck, even the bioanalyzer can be plus or minus 5-10bp.

Yes, this is likely the result of salt concentration differences between samples. [salt] impacts migration rate. A tailing reactions will have Mg++ and probably other salts as well as buffer.

--
Phillip

[seqgirl123]

Originally Posted by GW_OK

5. 18 cycles for PCR seems a bit high. Are you worried about PCR artifacts? You could probably get away with 10 or less. You only need ~2ul of a 10nM library in the end.
6. Going back to your shearing, are you really seeing genomic DNA after Covaris shearing?
7. Do you have access to a bioanalyzer? They're really helpful for this sort of thing.

thanks for the info.

5. 18 cycles is the number in the Ilumina protocol. Are you able to make genomic libraries in 10 cycles? I think for PE, that is the recommended from Illumina.
6. Yes, it is a genomic DNA sample that I am shearing. Some samples are not very pure and they contain RNA, but for most samples, they are genomic DNA
7. I could use a bioanalyzer and quantitate, but I was hoping someone here had already tried to reduce chimeric fragments by doing 2 sizeselections during a library prep or a sizeselection right after shearing so that I could have an idea of what kind of yield or losses to expect.

[GW_OK]

What version of the protocol are you using? All of the Illumina protocols I have call for 10-12 cycles.