scientific article; zbMATH DE number 3906284

From MaRDI portal
Publication:3683944





[For the entire collection see Zbl 0561.00017.] As an extension of the classical Moran model a founder population of haploid individuals with the two alleles a and A at a single locus is considered. Deaths occur randomly at discrete times. If the population consists of \(X_ n\) a-alleles and \(Y_ n\) A-alleles at the n-th death the \(n+1\)-th death is an a-allele with probability \(X_ n/(X_ n+Y_ n)\) and an A-allele with probability \(Y_ n/(X_ n+Y_ n)\). The dying individual is replaced by a random number \(U_ n\) of individuals where, conditioned on \(U_ n\) and \((X_ n Y_ n)\), the number of a-allele replacements is binomially distributed with parameters \(U_ n\), \(p_ n=X_ n/(X_ n+Y_ n).\) The \(\{U_ n\}\) are assumed as i.i.d. and independent of the \(\sigma\)-field \(F_ n\) generated by \(\{X_ k,Y_ k\); \(k\leq n\}.\) By investigation of the a-allele proportion process \(Z_ n=X_ n/(X_ n+Y_ n)\) it is proved that an ultimate loss of one of the allelic types is almost sure if \(P(U_ 1=1)=1\) and that there is a non-zero probability of ultimate polymorphism if \(P(U_ 1>1)>0\) (Theorem 1). However, if a selective advantage for one of the allelic types is introduced by assuming \(p_ n=(1+\sigma)X_ n/((1+\sigma)X_ n+Y_ n),\) \(\sigma >0\), ultimate loss of one of the allelic types is certain (Theorem 2). The proof is based on martingale convergence theorems, since \(\{Z_ n, F_ n\}\) is a martingale (Theorem 1) or a sub martingale (Theorem 2).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3683944)