I am a young student, just started working with protein mutations and I used NCBI info about different domains present in ITK. I used NCBI because it had much more data e.g. PI binding sites. However when I aligned those domains with my mutations, I was curious and checked those domain sequences with Uniprot. As presented below, there are some discrepancies between NCBI and Uniprot (I checked, it's the same Homo sapiens ITK). I wanted to ask if this is a common thing, why such data differ and which source is more reliable.
PH Domain according to Uniprot and NCBI:
I'm aware that the aminoacids at the beginning and the end of a domain are not as important as the ones in the middle, especially when special binding sites need to be formed like in this case. But saying that a mutation is within some domain or not, or even that given domain is 108? 104? aminoacids long seems inappropriate with different data from those online databases in mind. If this is normal, how do you deal with it?