AVID is a program for computing global alignments between sequences and for
ordering and orienting draft sequence.

The algorithm is described in:
Bray N, Dubchak I and Pachter L., AVID: A Global Alignment Program, 
Genome Research, Vol. 13, p 97--102 (2003).

Please cite this paper if you use AVID in your research.


The basic usage is as follows:

avid [options] file1 file2

where file1 and file2 are FASTA sequence files and the options to the program
are as described below.

IMPORTANT: If you are using AVID on Macintosh OS X, you must run the command
"unlimit stacksize" before running AVID.

In the case of draft sequence, file2 should be the name of the draft sequence
file. By default, the program expects that for each sequence file, there will
also be a masking file for that sequence located in the same directory. For
example, if the first sequence is named "human.fasta" then the program expects
there to be a file named "human.fasta.masked" which will be in the format of
a RepeatMasker .masked output(i.e. masked positions are indicated by Ns).


Options
-------

Running the program with the -r option causes it to reverse compliment the
second sequence before it computes the global alignment. This option has no
effect if the second sequence is draft.

The -nm={first,second,both} option specifies not to use masking for one or both
of the sequences. This is not recommended as not having the repeat locations
masked may cause the program to produce worse quality alignments.

The -obin option causes the program to use a binary format for it's output.
This format is described below.

The -opsl option causes the program to output a PSL file describing the
alignment.

Binary Output
-------------

The format for the alignment file is as follows: each byte of the file
corresponds to one line of the non-binary alignment file. The higher four bits
of the byte encodes the nucleotide from the base sequence and the lower four
bits encode the nucleotide from the second sequence. The encoding for each
nucleotide is gap = 0, A = 1, C = 2, T = 3, G = 4, N = 5.

The binary format for the info file is as follows. In the case of finished
sequence, the file has two lines:

[fasta line from the second sequence]
1 from to 0 0 0 0 {+,-} score secFrom secTo

where from and to represent the first and last locations of "significant"
alignment in the base sequence. Either + or - is output according to whether
or not the reverse option was specified. score is a score for the alignment
and, secFrom and secTo give the range of the cropped alignment in the second
sequence.

In the case of draft sequence, there will be two lines for every contig which
is placed by the program. Each pair of lines will look like this:

[FASTA line for this contig in the original sequence file]
n baseFrom baseTo mergedFrom mergedTo startChop endChop {+,-} score secFrom secTo

n is the index of this contig in the original sequence file. baseFrom and
baseTo give the range of this contig in the base sequence, after being cropped
until a significant alignment is found. If the contig is entirely gapped in
the base sequence then these number will both be 0. If no significant
alignment is found(the contig is entirely cropped) then these numbers will
both be -1. mergedFrom to mergedTo give the range of
this contig in the merged draft sequence. Due to overlapping between
consecutive contigs, this range may be shorter than the original contig.
startChop and endChop give us the information about the overlapping of
consecutive contigs. For example, if our file contained the lines:

D002.fasta.screen.Contig47 (90178-99052, plus)
47 219333 226457 91522 99146 0 62 + 5143
D002.fasta.screen.Contig50 (97006-96802, plus)
50 226458 234292 99246 107077 1124 0 + 4872

Then contig 47 has endChop = 62 and contig 50 has startChop = 1124. This means
that there was a total overlap of 62 + 1124 = 1186 between these two contigs
and that the contigs were merged by chopping 62 bases off the end of contig 47
and 1124 off the beginning of contig 50.(startChop and endChop are chosen with
the goal of having the best quality merged sequence)

Finally, the line ends with + or - depending on whether the contig was oriented
forward or reverse complement, respectively.