How to calculate percentage identity/error rate from local alignment score?

Question

[crivenster]

Hi,

I have a local alignment on similar sequences using Smith-Waterman algorithm. I have alignment scores generated,how to calculate the percentage identity between the sequences? can it be calculated from the alignment score?

Thanks.

[lh3]

Unfortunately, it is not possible to precisely compute the identity merely from a score.

[mikhail.shugay]

What software have you used to calculate the alignment?

[crivenster]

I have written a small piece of code, implemented Smith-Waterman algorithm in java as i had some issues with Biojava package. But then i wanted to also find the percentage identity between the sequences. The online searches dint give a clear or a definite picture on calculating the percentage identity between two nucleotide sequences.