Performance comparison between bootstrap and multiscale bootstrap for assessing phylogenetic tree for RNA polymerase
Phylogenetic inference refers to the reconstruction of evolutionary relationships among various species that is usually presented in the form of a tree. This study constructs the phylogenetic tree by using a novel distance-based method known as Modified one step M-estimator (MOM) method. The branche...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Universiti Kebangsaan Malaysia
2015
|
| Online Access: | http://journalarticle.ukm.my/9374/1/15_Safinah_.pdf http://journalarticle.ukm.my/9374/ http://www.ukm.my/jsm/malay_journals/jilid44bil11_2015/KandunganJilid44Bil11_2015.html |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Phylogenetic inference refers to the reconstruction of evolutionary relationships among various species that is usually presented in the form of a tree. This study constructs the phylogenetic tree by using a novel distance-based method known as Modified one step M-estimator (MOM) method. The branches of the phylogenetic tree constructed were then evaluated to see their reliability. The performance of the reliability was then compared between the p-value of multiscale bootstrap (AU value) and bootstrap p-value (BP value). The aim of this study was to compare the performance between the AU value and BP value for assessing phylogenetic tree of RNA polymerase. The results have shown that multiscale bootstrap analysis can detect high sampling errors but not in bootstrap analysis. To overcome this problem, the multiscale bootstrap analysis has reduced the sampling error by increasing the number of replications. The clusters were indicated as significant if AU values or BP values were 95% or higher. From the analysis, the results showed that the BP and AU values differ at 11th and 15th branch of the phylogenetic tree. The BP values at both branches were 72 and 85%, respectively, thereby making the cluster not significant but by looking at the AU values, the two branches were more than 95% and the clusters were significant. This was due to the biasness in calculation of the probability of bootstrap analysis, therefore, the multiscale bootstrap analysis has improved the calculation of the probability value compared to the bootstrap analysis. |
|---|