If I am considering Taq polymerase to PCR amplifying 1.3kb of 7.5kb genome, knowing that the mutation rate of Taq is 1 base in 10000 bases (it can be translated in this context as 1 base every ~8 amplicons), how to calculate the mutation profile after 40 cycles of PCR?
Since the occurrence of mutation is a random process, the outcomes for a series of identical PCR reactions will all be different. A very important factor is WHEN a particular mutation occurs. If a mutation occurs during a very early cycle, it will be represented in a large proportion of the final product. But if it occurs at a late cycle, when the replication efficiency is no longer high, it may be hard to detect. If you need your product to be mutation free, you should consider using one of the many polymerases with greatly reduced error rates. To insure the absence of any copy error effect, you can then either sequence the product, or use multiple independent replicate products to proceed further in your experiment.
Thank you very much David,
I am actually simulating a system, and I am wondering how to calculate the mutation profile given the conditions in the question above, so I would roughly be able simulate similar frequency and nature of this mutation in the library? My question is about a statistical/mathematical method that can give me a rough estimation?
Many thanks
another complication is that the figure of 1/10,000 is based on extension at 72c. David is correct about the importance of when the mutation happens but additionally when in each cycle is relevant. Primers anneal at lower than 72c and Taq makes more errors at lower temperatures but it still extends at all temperatures so I would expect more errors near to the primers when annealing has just taken place but extension quality is poor. Also while most of the errors are random I do think that some sequences are moreprone to replication errors than others ( CA repeats and poly A for example ) and possibly CG ( some muations are more common than others in the genome ) so it may be a difficult problem
Yes, Paul has a good point. If you would do an infinite number of replicates of the same PCR reaction, then the profile of the mutations present in the pool of all of the infinite number of separate products would reflect the sequence-specific (i.e. sequence-dependent) likelihood of a mutation occurring at each position. I would not care to hazard a guess as to what finite number of reactions would need to be done to get a good approximation to the infinite reaction result .
simplistically and making assumptions about the uniformity of mutation production a very rough calculation of the number of mutations comes from the basic pcr equation y=2expx where duplication at each pcr step then x is the number of cycles . This can only be used in the exponential phase then y is the amount of product compared with the starting amount and assuming perfect . So if the starting number of target molecules ( molecules of dna x number of copies per genome) is M and the number of bases between the 3' ends of the primers is N then the number of bases after x cycles is mN2expX so the number of mutations in your example is simply
mN2expX/10,000 but this is very approximate and only true for the early cycles in a perfect PCR and certainly not true for 40 cycles where the later cycles are very inefficient