Dear Said, 

Bootstrap support values must be analyzed carefully. There is much debate about the value that indicates a statistically well-supported grouping. Most researchers consider 70% or above as a good support, but others consider as low as 50% as probably significant. However, you may obtain a higher support if you include more information in your phylogenetic analyses, i.e. more loci or a longer fragment of the same gene.

One interesting thing is that sometimes you can get bootstrap values below 70%, but when you analyze the same dataset with a different method, for example Bayesian analysis, you get clearly good statistical support (posterior probabilities of 0.95 or higher).

Success with your research!

Hugo

In general, the ML bootstrap value greater than 80% is supposed to be good. The value 63% maybe a little low.

You can find your answers in this article of Hillis and Bull who simulate bootstrap values under some assumptions.

http://sysbio.oxfordjournals.org/content/42/2/182.short

And in this topic too https://www.researchgate.net/post/How_can_I_interpret_bootstrap_values_on_phylogenetic_trees_built_with_Maximum_Likelihood_method/1

Dear Said, 

Bootstrap support values must be analyzed carefully. There is much debate about the value that indicates a statistically well-supported grouping. Most researchers consider 70% or above as a good support, but others consider as low as 50% as probably significant. However, you may obtain a higher support if you include more information in your phylogenetic analyses, i.e. more loci or a longer fragment of the same gene.

One interesting thing is that sometimes you can get bootstrap values below 70%, but when you analyze the same dataset with a different method, for example Bayesian analysis, you get clearly good statistical support (posterior probabilities of 0.95 or higher).

Success with your research!

Hugo