[ -- Internship Proposals -- ]

Background

ThermoFisher has a PCR Calculator, which is basic. This just calculates ~ [1-l x numcycles x u] (l, fragment length; u, error rate), which is nothing else than the probality for a molecule of length l x numcycles to have no errors. Thus, it assumes 100% efficiency and ignores that some molecules are not replicated and contribute to the overall number of molecules without errors.
Here, I use a more rigourous approach (a good article can be found here ). Just enter the parameters below and submit the code (using the Opencpu cloud server; this might takes a few seconds to get the result). Also notice that the calculations are made for 1E3 molecules only.

Enzyme

Taq Q5 Phusion Deep Vent Pfu PrimeSTAR KOD Kapa HIFI DV Exo-

Length

100 200 400 800 1000 1500 2000 3000 5000

Cycles

15 20 22 25

#PCR Fidelity Calculator (R Code) # Adapted from doi.org/10.3929/ethz-a-006088024 # W. Grange (2018) #Declare Variables # Rate # u<-1.5E-4 # Fragment length # l<- 2000 # Number of Cycles # numcycles=25 # Number of mutations to calculate mut<-10 # Number of molecules N<-1000 #Efficiency (doi.org/10.1093/nar/25.15.3082) lambda<-c(rep(0.872,20),rep(0.743,5),rep(0.146,5)) #Init data<-data.frame(matrix(0,ncol = mut , nrow = (numcycles+1))) vec_p<-rep(0,mut) vec_p[1]<-N data[1,]<-vec_p #Shift Function (https://stackoverflow.com/questions/26997586/r-shifting-a-vector) shift <- function(x, n, invert=FALSE, default=0){ stopifnot(length(x)>=n) if(n==0){ return(x) } n <- ifelse(invert, n*(-1), n) return(c(rep(default, n), x[seq_len(length(x)-n)])) } #Calculate mutation prob pr_mutation<-sapply(1:mut, function(i){dmultinom(c((i-1),l-i+1),l,prob=c(u,1-u))}) #First cycle (could be in the loop too) #Determine vector of mutations n<-round(N*lambda[1]) vec<-round(pr_mutation*n) #Check that round number is ok if (sum(vec) != n){ imax<-which.max(abs(round(pr_mutation*n)-pr_mutation*n)) if (sum(vec) > n){ vec[imax]<-vec[imax]-1 } else if (sum(vec) < n){ vec[imax]<-vec[imax]+1 } } #Add previous vector vec<-vec+vec_p molecules<-vec data[2,]<-molecules #Cycles for (k in 1:(numcycles-1)) { # molecules has "mut" elements (molecules have 0, 1, 2 ... mutations) # Calculate number of new molecules generated n<-round(sum(molecules)*lambda[k+1]) # For these new molecules, calculate the number of molecules with 0, 1, 2 .... mutations vec<-round(pr_mutation*n) # Check that round number is ok if (sum(vec) != n){ imax<-which.max(abs(round(pr_mutation*n)-pr_mutation*n)) if (sum(vec) > n){ molecules[imax]<- molecules[imax]-1 } else if (sum(vec) < n){ molecules[imax]<- molecules[imax]+1 } } # multinomial distribution prob_multi<-sapply(1:mut,function(i){molecules[i]/sum(molecules)}) Error<-data.frame(matrix(0,ncol = mut , nrow = mut)) Error_shifted<-data.frame(matrix(0,ncol = mut , nrow = mut)) Error[1,]<-rmultinom(1,size=vec[1],prob=prob_multi) Error_shifted[1,]<-Error[1,] for (i in 2:mut) { Error[i,]<-rmultinom(1,size=vec[i],prob=prob_multi) Error_shifted[i,]<-shift(Error[i,],(i-1)) } SUM<-apply(Error_shifted,2,sum) # Update Molecules molecules<-molecules+SUM data[(k+2),]<-molecules } res<-signif(100*molecules/sum(molecules),3) paste("[Code still under test] % molecules with 0 error:", res[1], ', 1 error: ', res[2], ', 2 errors: ', res[3])
Submit Code

.../...