I want to identify if certain phosphorylation sites are conserved for protein X across humans and yeast. I know from MS data that there are 4 phosphorylation sites in Human protein X. In order to identify if these residues, and therefore potentially the phosphorylation sites, are conserved, is it correct to align both sequences using pairwise comparisons by using software like Needle? Or should I rather perform a bigger alignment and use clustal omega? Are there better approaches?
I'd perform a bigger alignment using clustal or mafft or something. Pairwise alignments will give you a lot of noise from random chance and the set of pairwise alignments will be really large. Multiple sequence alignment will detect smaller conservation signals.
Compare against several species of yeast and maybe other mammals for context. When the alignment is done, perhaps draw it? I like logos for vertically large alignments.