Hi Narjes,

In order to generate a reliable tree, you need a very reliable multiple sequence alignment.

I’d also suggest a smaller set to begin with (max 100 sequences), so you can compare between your preferred ML method and bayesian (e.g. MrBayes).

You may try to reduce redundancy of very similar sequences using CDHIT.

After constructing a well-resolved tree from the manually-curated MSA 100 sequences, you may try an ML method (e.g. FastTree) applied on your 1,148 sequences and compare your results with your well-resolved, small tree, since fast methods that are capable of such big trees should be considered with care.

 Thank you Eyal 

In fact I am beginner, I've constructed my phylogenetic  tree for my original data and it seems to be good, I am working on such fungal protein which will be considered  later as antifungal . As we know that high similarity between human protein and the fungal one swill be problematic so I have to align all these protein sequences and constructing tree which might help in identifying specific protein with low similarity or even they are specific to fungi. I failed to construct my tree using fast tree and instead i used maximum likelihood with post rap 500.

Mr Bayes  as I understood it is good but i couldn't download it .

If you are having trouble downloading programs or do not have access to computing clusters you can always use CIPRES, which is an online gateway with several phylogenetic programs (BEAST, Mr. Bayes, RAxML, etc.). It is easy to use and a great resource for researchers who do not have access to a high performance computing clusters. Hope this helps.

https://www.phylo.org/portal2/login!input.action

Thank you Max, I will try now