What is the most efficient way of multiple sequence alignment (MSA) to get new MSA based on k+1 from based on k sequence?

When performing a BLAST how can we guarantee that our hits are reasonable and adapted answer to our query sequence?

What tool is the most efficient to determine evolutionary distance among protein sequences? 
Additional Requirements 

Level of Detail: Only answer needed