Following up annotation, starting with the Rice Genome Program at http://rgp.dna.affrc.go.jp/cgi-bin/statusdb/status.pl , it looks as though answering this would almost answer every question we could come up with about genome sequencing. Now that we have a figure, 22%, on what we are doing, why not tell us a bit about what ISB is doing? Starting with getting work units to us and continuing with what is being done once the results are returned. We only have part of the picture. If you just blog a bit about what you have to do day by day, we will eventually be able to ask pertinent questions to round out our picture.

Ok this is a bit over the top, but to see the latest publication
about how we use Rosetta at the ISB to get protein function
leads see:

http://genomebiology.com/2004/5/8/r52

this is a research pub, and a lot technical... BUT
look first at figure 1. This is effectively the pipe we
use to identify sequences to fold, and it consists of the
following steps:

1. use sequence based methods to figure out
posible functions for as many proteins as posible.
[word-word / sequence/sequence] matches between
protein your interested in and known protein.

2. If no word match for a protein of part/domain of a
protein then pass that protein or protein domain to Rosetta
and/or other structure prediction techniques.

3. Take those proteins and put them on the grid for
grid_Rosetta to chew on.

4. distill function from structures by (among other things)
looking for structure-structure matches [structural
similarity is conserved across a greater evolutionary
distance that sequence similarity]

1+2 result in fragments of local structure
3 puts those fragments of structure together to
make global folds
4 attempts to distill useful info from the folds

Ok here is an example of what an annotation is and how
we can get it from the grid. Just one example of an annotation
and how one arives at it with structure as a piece of evidence.

1. protein is identified as important for any number of reasons:
eg. my biologist buddy, Nitin Baliga for instance, tells me that
a protein (VNG1462H) is something the cell makes more of
when it is stressed, and that the other proteins the cell makes
at these times include b12 synthesis enzymes.

2. I predict the structure (on the grid + ISB/UW pre and post-process pipeline).

3. We see what the structure mathches and see that
it matches Haem (b12 is a haem ring) binding folds similar to folds seen
in the b12 biosynthetic pathway.

4. we combine our two pieces of evidence to form
a hypothosis that VNG1462H carries out step X in the
colbamine(b12) biosynthetic pathway.

5. repeat, until whole halobacterium genome is understood
via a combination of computational (structure+sequence matching)
and experiemental (gene arrays and proteomics and genetics, etc.)
methods.