SEQanswers

Go Back   SEQanswers > Applications Forums > Sample Prep / Library Generation



Similar Threads
Thread Thread Starter Forum Replies Last Post
ATAC-seq cutting frequency analysis cloverliang Bioinformatics 0 08-26-2015 08:00 AM
Cutting Truseq Adapaters Trimmomatic arcolombo698 Bioinformatics 1 02-19-2014 01:28 PM
Cutting of Double Stranded Clusters on Flowcell? NextGenSeq Illumina/Solexa 4 09-03-2013 07:21 AM
Cutting corners with Illumina Sample Prep Rocketknight Sample Prep / Library Generation 21 03-29-2013 10:17 AM
TopHat: cutting reads into segments nike00 RNA Sequencing 2 07-15-2011 07:30 AM

Reply
 
Thread Tools
Old 11-05-2015, 01:26 PM   #1
Carcharodon
Member
 
Location: Honolulu, HI

Join Date: Jul 2015
Posts: 32
Default Rare cutter not cutting? (ddRAD)

Greetings, everyone!

I've been testing my restriction enzymes, and have candidates for a rare cutter and a common cutter. The rare cutter is BamHI, and my frequent/common cutter is PstI.

I tested each enzyme separately and together (B+P = BamHI + PstI). Each digestion was allowed to proceed for ~ 6 hours, and reaction conditions were set according to the manufacturer's recommendations.

I ran these on gels next to undigested genomic DNA; I've attached the results below.

My concern is that there is still a band present in the BamHI lane, though it appears in both cases that it is slightly smaller than the undigested genomic DNA (the ladder is Hyperladder I; sorry for the over-exposure).

Is this especially problematic, or is this something that is commonly seen when performing these types of digestions with "rare" cutters?

I will note that I am aiming for a highly reduced-representation library; my organism's genome is ~ 8 Gb in size, and I'm trying to maximize coverage - so an enzyme that doesn't cut very often is desirable. But this has me wondering whether this one is cutting too infrequently (if at all)?

Many thanks,
Sean
Attached Images
File Type: jpg SJC_RE_07-23-2015.jpg (85.2 KB, 22 views)
File Type: jpg SJC_RE_07-24-2015.jpg (90.8 KB, 17 views)
Carcharodon is offline   Reply With Quote
Old 11-05-2015, 03:27 PM   #2
nucacidhunter
Jafar Jabbari
 
Location: Melbourne

Join Date: Jan 2013
Posts: 1,210
Default

It seems that it is cutting. If you are doing ddRAD, you could ligate adapters and do a PCR. Amplification will be a good indication of how it has worked.
nucacidhunter is offline   Reply With Quote
Old 11-06-2015, 06:47 AM   #3
ATϟGC
Member
 
Location: Canada

Join Date: Jun 2013
Posts: 44
Default

In theory, both BamHI (GGATCC) and PstI (CTGCAG) should cut with equal frequency in any given genome as they both have a 6bp recognition sequence with the same frequency of A-T and G-C base pairs (33.3% and 66.6%, respectively). The exception would be if one 6 bp motif is far more frequent than the other. For example, if it is sequence in a repetitive element.

It is possible that your BamHI reaction did not work properly. I would not assume that the slight shift in apparent size is due to cutting as different quantities of DNA and remaining enzyme-bound DNA can cause apparent size differences that do not really exist.

You may want to do an in silico digestion with BamHI and PstI if you have not tried already and have genome sequence data of your focal species or a close relative available. I have found ddSilico very easy to use (see additional file 1 here: http://www.biomedcentral.com/1471-21...275/additional)

If your genome is more AT than GC rich and want to stick with 6bp cutters you may want to give NsiI (ATGCAT)a try as it is 66.6% AT. Since NsiI will leave the same overhang sequence as PstI, you may want to avoid using the two together to avoid problems in library preparation.

I would add one caveat about using PCR to test if BamHI has worked: If you do not eliminate fragments with two PstI ends from the B+P sample you may still get amplification of fragments of the appropriate size that are essentially useless for high-throughput sequencing as they will not have the appropriate adapters on each end.
ATϟGC is offline   Reply With Quote
Old 11-06-2015, 09:39 AM   #4
nucacidhunter
Jafar Jabbari
 
Location: Melbourne

Join Date: Jan 2013
Posts: 1,210
Default

That theory is incorrect because order of bases in a genome is not random. Two genome with similar size and GC% can have different number of RAD fragments.

In ddRAD 1st cycle of PCR is primed from rare cuter adapter and fragments ligated to frequent cutter adapter in both ends can not be amplified.
nucacidhunter is offline   Reply With Quote
Old 11-06-2015, 10:02 AM   #5
Carcharodon
Member
 
Location: Honolulu, HI

Join Date: Jul 2015
Posts: 32
Default

Thanks so much for your replies!

Quote:
Originally Posted by ATϟGC View Post
In theory, both BamHI (GGATCC) and PstI (CTGCAG) should cut with equal frequency in any given genome as they both have a 6bp recognition sequence with the same frequency of A-T and G-C base pairs (33.3% and 66.6%, respectively). The exception would be if one 6 bp motif is far more frequent than the other. For example, if it is sequence in a repetitive element.

It is possible that your BamHI reaction did not work properly. I would not assume that the slight shift in apparent size is due to cutting as different quantities of DNA and remaining enzyme-bound DNA can cause apparent size differences that do not really exist.

You may want to do an in silico digestion with BamHI and PstI if you have not tried already and have genome sequence data of your focal species or a close relative available. I have found ddSilico very easy to use (see additional file 1 here: http://www.biomedcentral.com/1471-21...275/additional)
I've been using SimRAD (an in-silico digestion package for R). If memory serves, my simulated genomes resulted in little difference in cutting frequency between the two enzymes.

However, using the whale shark genome as a proxy (I'm working with a shark from the same superorder...it was as close as I could get!), it appears that PstI cuts about 4.5 times more often than BamHI on average.

Data from digests performed on human genomes seem to generally support the idea that PstI is a more frequent cutter:

GGATCC (BamHI) = 126.1 cut-sites/Mbp
CTGCAG (PstI) = 452.8 cut-sites/Mbp

http://tools.neb.com/~posfai/TheoFra...est.human.html

Pretty interesting stuff!

Just to be safe, I've contacted the manufacturer and they are sending a sample replacement. Another test should ease my worries.

Quote:
Originally Posted by ATϟGC
I would add one caveat about using PCR to test if BamHI has worked: If you do not eliminate fragments with two PstI ends from the B+P sample you may still get amplification of fragments of the appropriate size that are essentially useless for high-throughput sequencing as they will not have the appropriate adapters on each end.
This is true, although according to my simulations, and based on my adapter/oligo design, only fragments with cutsites from BamHI should be amplified. So there will be a few BamHI/BamHI fragments, but they should (in theory) be in the minority, if BamHI is indeed a rare cutter.

SimRAD simulations with the whale shark draft genome seem to indicate that these fragments will make up 10% - 15% of my total number/volume, but I'm hesitant to trust these simulations quite so explicitly.

Thanks again for your responses. This forum is a wonderful resource, and I appreciate all of your advice! I'll do my best to pay it forward.
Carcharodon is offline   Reply With Quote
Old 11-06-2015, 11:24 AM   #6
ATϟGC
Member
 
Location: Canada

Join Date: Jun 2013
Posts: 44
Default

You are correct nucacidhunter that the theory does not always hold true. I apologize that my second sentence did not clearly convey that it is an approximation based on the assumption that base order is completely random, which obviously is not true. I also should have mentioned that I was assuming that PstI was the P1 end (rare cutter) for the first round of PCR when in this case it should be BamHI since it does appear to cut less often. I know that Peterson et al (2012) warn of potential P1-P1 amplification products so I thought I would mention it as a possibility.

Sean, your evidence certainly suggests that PstI is cutting much more frequently than BamHI, so what you see on the gel is likely the real deal.
ATϟGC is offline   Reply With Quote
Old 11-06-2015, 05:11 PM   #7
nucacidhunter
Jafar Jabbari
 
Location: Melbourne

Join Date: Jan 2013
Posts: 1,210
Default

Quote:
Originally Posted by Carcharodon View Post
SimRAD simulations with the whale shark draft genome seem to indicate that these fragments will make up 10% - 15% of my total number/volume, but I'm hesitant to trust these simulations quite so explicitly.
From those 10-15% only a small subset would be in size-selection window and end up in ddRAD library. In my experience they are below 3% although it would vary from genome to genome and the enzymes used for digestion. They can cause overestimation of library concentration by qPCR resulting in slightly lower cluster number than expectation.

If you suspect that your batch of enzyme has lost activity you can digest another species DNA such as E coli with it for testing.

I have had species that like yours did not seem has been digested and I could not see any fragment below 1kb on Bioanalyser HS DNA Chip (prominently a large peak at higher molecular weight) but after double digest and size-selection in a range below 400 bp, PCR yielded good libraries and results were fine.
nucacidhunter is offline   Reply With Quote
Old 11-11-2015, 07:03 AM   #8
pmiguel
Senior Member
 
Location: Purdue University, West Lafayette, Indiana

Join Date: Aug 2008
Posts: 2,307
Default

Are you sure you didn't switch the samples during loading? I mean an 8GB genome, that is probably a plant. Plants with genomes that large are massively cytosine-methylated. PstI is blocked by CNG methylation (but not by CG methylation!)
So, anyway, the lane you labeled "BamHI" looks like typical large genome plant genomic DNA digest with PstI.
Of course this may not be plant DNA. In which case, never mind...
--
Phillip
pmiguel is offline   Reply With Quote
Old 11-11-2015, 09:55 AM   #9
SNPsaurus
Registered Vendor
 
Location: Eugene, OR

Join Date: May 2013
Posts: 465
Default

Check out the username of the OP, pmiguel.

and http://www.genomesize.com/result_species.php?id=1701
__________________
Providing nextRAD genotyping and PacBio sequencing services. http://snpsaurus.com
SNPsaurus is offline   Reply With Quote
Old 11-11-2015, 10:13 AM   #10
pmiguel
Senior Member
 
Location: Purdue University, West Lafayette, Indiana

Join Date: Aug 2008
Posts: 2,307
Default

Quote:
Originally Posted by SNPsaurus View Post
Check out the username of the OP, pmiguel.

and http://www.genomesize.com/result_species.php?id=1701
Ah.

--
Phillip
pmiguel is offline   Reply With Quote
Reply

Tags
ddrad, library preparation, restriction enzymes

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off




All times are GMT -8. The time now is 09:24 PM.


Powered by vBulletin® Version 3.8.9
Copyright ©2000 - 2019, vBulletin Solutions, Inc.
Single Sign On provided by vBSSO