ABI Traces version 2

Andy Law Big Nose Andy.Law at bbsrc.ac.uk
Mon May 22 08:09:01 EST 1995


Anyone got any further with sorting out the changes between versions 1 and 2?

I have the following so far (these are only informed guesses. I don't have
the version 2 software to test out what I think changes in the file will
do to the analysis)

The offset at the start of the file that used to point to the AUTO data
structure no longer points there. In the two files I have seen, the offset
points beyond the end of the file. Presumably there is some magical
formula that can be applied to the offset, but I haven't found it. Brute
force suggests reading the file in and scanning for the 'AUTO' string (see
later) I believe that ABI used to read in the whole file into memory in
version 1 so that won't influence their practices.

There are several new data structures. These are:-

   'ADSP', 'AEPt' (Nos 1 and 2), 'ASPF', 'ASPt' (Nos 1 and 2), 
   'B1Pt' (Nos 1 and 2) & 'drty'

Since the data structures are in alphabetical order, the 'AUTO' structure
is no longer the first in the sequence.

In addition, there are no 'GELT', 'LIMC' or 'SMOD' structures.

There are one extra of each of the 'PDMF', 'PPOS' and 'SPAC' structures.

There is one fewer 'SVER' structure.

Given that the 'LIMC' structure used to point to the data pertaining to
the Base Call Base 1, Base Call Start and Base Call End data, I imagine
that this data is now stored directly in the 'B1Pt', 'ASPt' and 'AEPt'
structures respectively.

I also imagine that the 'drty' structure is a flag structure indicating
the 'dirty' status of the file?

That just leaves the 'ADSP' and the 'ASPF' structures. Any clues, anyone.

I haven't looked into how the base calls and peak locations differ between
the two versions.
-- 
Andy Law

( Andy.Law @ bbsrc.ac.uk )
( Big Nose in Edinburgh )



More information about the Staden mailing list