How to get enough product from limited samples without generating mutations?
Posted 23 March 2010 - 04:04 AM
Now I have some genomic DNA sample from a carrier with proviral genome (HTLV) integrated into his genome. I want to sequence some specific region of the proviral genome, which may be several kb in length. Normally, for a 2kb region, I use a polymerase with high fidelity to amplify this region in less than 30 cycles, which had been proved to generate no mutations in other experiments. Because in carrier stage, proviral genome may have different copies, if I sequence the PCR product directly, I will only get the sequence of the predominant copy. So I clone this PCR product (which should be a mixture of products from different proviral genome copies) into TOPO vector, transform E.coli, and then pick like 20 or more colonies and sequence each clone. Then I could make an alignment of sequences from all the clones and compare.
To reach a good efficiency in blunt-end cloning, I need enough PCR product. However, I cannot increase the amount by simply increasing the cycles, because that will generate mutations. So I have only to increase the total amount of PCR reactions, usually I used 4X 50 ul to get enough product. But now I dont think I have enough templates for PCR amplification. That is the problem.
Does anybody have such experience? I really need your help. Thanks in advance.
Posted 23 March 2010 - 04:43 AM
Posted 23 March 2010 - 11:14 AM
For example when using Phusion you have 2.6% products with single mutation at 30 cycles vs. 3.5% at 40 cycles. No real difference.
Posted 23 March 2010 - 05:50 PM
Maybe I didnt make it very clear. The objective is to check the mutations of proviral genome in carriers, so the sequence I got from transformants are supposed to contain mutation(s). If PCR also generated mutations, it would be impossible to tell if the mutations really exist in carriers.
If I change my question like this: Sequences in carriers contain mutations, but probably in a percent lower than the error rate of PCR enzymes. In this case, how could I get enough product for sequencing and avoid the possible interference from PCR enzymes?
I use Herculase II combo from Stratagene. It works well for the cloning and sequencing of 2kb.
Posted 23 March 2010 - 06:01 PM
Posted 23 March 2010 - 07:17 PM
Thank you for your reply. but I didnot get it. do you mean that if I only make 20 clones, there is an extremely low chance to pick the clone with pcr-generated mutations?
Yes, I understand. You could quantify your mutation rate by amplifying something of known sequence, if you think this is a real issue. I don't. I also think you vastly overestimate the amount of insert you need to make thousands of clones.
Posted 24 March 2010 - 01:06 AM
Posted 24 March 2010 - 05:57 AM
Also, if you are concerned with PCR induced mutations, you could turn to single molecule sequencing that doesn't require amplification (or can work with limited amount of PCR product).
I know it's hard to get your hands on such a machine but, if you could find a collaborator, that would be the best option I think. You'd have thousands and thousands of "clones" instead of 20.
Posted 25 March 2010 - 07:24 AM
about mutation rate of carriers and error rate of PCR, I have ever done this before. generally, using the method metioned above, I found a mutation in several clones from total 20 clones. to understand if this mutation real exists or was just caused by polymerase, I amplified the sequence with the mutation by the same method, cloned it, picked 20 clones and sequenced all of them. I got no new mutations this time. so I think this method or the polymerase I used wouldnot generate any false positive result. do you think its right?
I know 20 is not a big number, but depending on what we have now, I cannot do more. I also discussed the possiblity using next-generation sequencing with other people, but this is not a big project and my boss wouldnot agree. so what I can do is only optimize the method and make it more efficiency. right now, I just found that I could use much less insert than before in the TOPO cloning. if that finally works, half of this problem will be solved. just the mutations I found, would be a real problem.