Comparison of statistical methods for genetic similarity evaluation of maize inbred lines

Abstract

Conventional breeding methods have been aided by molecular genetic techniques giving the chance for efficient improvement in creation of maize hybrids. Proper choice of statistical methods for data analysis is very important because it ensures greater reliability. The aim of this study was to determine the most suitable statistical approach for molecular marker data analysis. SSR markers were used for the analysis of 10 maize inbreds. Genetic similarity/distance was calculated using three types of data: binary, allele frequency based on densitometry and allele frequency according to band size data applying Simple matching, Jaccard's and Rogers' coefficient. Cluster analysis was performed in NTSYS, 2.11a software. The highest value for Spearman's rank of correlation (0.95) was detected between distance matrices based on binary data. The results showed that binary data Jaccard's coefficient) and allele frequency data based on fragment sizes (Rogers' coefficient) gave identical clusters by visual inspection and according to CIc index.

Metode klasične selekcije kukuruza se dopunjuju tehnikama molekularne genetike u cilju efikasnijeg dobijanja pouzdanih rezultata, pri čemu pristupi u obradi podataka imaju veliki značaj u ostvarivanju ovog cilja. Primenom SSR molekularnih markera analizirano je 10 samooplodnih linija kukuruza. Različite statističke metode su upoređene sa ciljem da se utvrdi najpogodnija za određivanje genetičke srodnosti ispitivanih genotipova. Genetička sličnost/distanca je izračunata korišćenjem tri tipa podataka: binarni podaci (1,0), frekvencija alela izračunata pomoću denzitometrije i frekvencija alela izračunata na osnovu veličina umnoženih fragmenata u baznim parovima primenom Simple matching, Jaccard i Rogers koeficijenata. Klaster analiza je urađena u NTSYS, 2.11a softveru. Najveća vrednost Spirmanovog koeficijenta (0.95) je utvrđena između matrica genetičkih sličnosti/distanci izračunatih na osnovu binarnih podataka. Rezultati ukazuju da se identični dendrogrami dobijaju korišćenjem Jaccard-ovog koeficijenta izračunatog za binarni zapis i Rogers-ovog koeficijenta izračunatog na osnovu frekvencija alela određenih prema veličini umnoženih fragmenata, kako vizuelnom ocenom tako i na osnovu CIc indeksa.