Model indel sequencing errors dependent on neighbour base qualities?

[krakau]

Hi all,

I am wondering in how far indel sequencing error probabilities are reflected in the base qualities of the neighbour bases in Illumina reads.
Does anyone know, if there exist any tools which model those possible indel errors for aligning or snp/indel calling?
I know e.g. about LAST, which can model different substitution sequencing errors, but as far as I know it doesn't take indel errors into account.

For an insertion error it's maybe enough to use the assigned quality, but what would be a good aproach for deletions?
Maybe using qualities of both neighbours or only of bases behind the deletion?

Thanks in advance.