Grouping of Hamedan garlic clones and determination of effective traits on in vitro bulblet production using multivariate statistical methods

Garlic has the second rank in Allium species in the case of highly used plant species after onion with high nutritional and medicinal values. Hamedan province is one of the important areas of garlic production in Iran. Meristem culture is an in vitro bulblet production method for removing the viral diseases. Multivariate statistical methods simultaneously evaluated several genotypes in terms of numerous characters and are widely used to assess the genetic diversity. The purposes of current study are use of multivariate methods to assess and initial classification of Hamedan garlic clones according to micropropagation and in vitro bulblet production traits derived from meristem culture and preliminary identification of most desirable traits on bulblet production.
Material and Ten different garlic clones of Hamedan province were used in this experiment as plant materials. The garlic cortex were separated and washed by distillated water then sterilized by ethanol (70%) for 10 min and sodium hypochlorite (2%) for 30 min.. After decontamination, meristems were separated in sterile condition under a binocular microscope and were cultured on MS culture medium supplemented with 5 µM NAA and 10 µM BA. The cultured samples were then transferred to growth chamber with 25ºC temperature and photoperiod of 16/8 h (day/night). Meristem culture experiment was carried out as a completely randomized design with 3 replications in 2015. After 38 days, micropropagation and bulblet production traits were measured on different clones of garlic. Finally, multivariate statistic methods were used to classification of clones and to detection of the most effective traits on in vitro bulblet production.
Results of principal components analysis showed that three first principal components explained 71 % of the total variance. According to these results, bulblet globularity, root weight, leaf number, mean of leave length, bulblet number and the longest leaf showed highest effects on two first principal components. Stepwise regression analysis indicated that leaf number, bulblet globularity and mean of leave length were the most important effective traits on yield (bulblet number) and explained 77.88 % of total variance. Based on path analysis results, the traits of number of leaf and mean of leave length showed the maximum positive direct and significant at p