filterBam - filter BAM file for use with AUGUSTUS tools
filterBam --in in.bam --out out.bam [options]
The input file must be sorted lexicographically by 'queryname',
with e.g.
•sort -k 1,1 [be aware: 'export LC_ALL=C'
might be used because sort ignores characters like ':'] Also, please bear in
mind that this will require converting your BAM file into SAM.
•samtools and bamtools provide
facilities to do the sorting, but they are not guaranteed to work because of
the problem mentioned above.
•In the case of samtools, the command is:
'samtools sort [-n] file.bam'. The option [-n] should sort by query name, just
as 'sort -k 10,10' would do in a PSL file. Without options, the sorting will
be done by reference name and target coordinate, just as a 'sort -n -k 16,16 |
sort -k 14,14' would do with PSL. For more information check the man page
included in samtools distribution.
•bamtools can also sort bam files: bamtools
sort -queryname -in file.bam, but only provides the option to do it by
queryname.
If the option 'paired' is used, then alignment names must include
suffixes /1,/2 or /f,/r.
--best
output all best matches that satisfy minId and minCover
(default 0)
--noIntrons
do not allow longer gaps -for RNA-RNA alignments-
(default 0)
--paired
require that paired reads are on opposite strands of same
target (default 0). NOTE: see prerequisite section above.
--uniq
take only best match, iff, second best is much worse
(default 0)
--verbose
output debugging info (default 0)
--insertLimit n
maximum assumed size of inserts (default 10)
--maxIntronLen n
maximal separation of paired reads (default 500000)
--maxSortesTest n
maximal sortedness (default 100000)
--minCover n
minimal percentage of coverage of the query read (default
80)
--minId n
minimal percentage of identity (default 92)
--minIntronLen n
minimal intron length (default 35)
--uniqThresh n
threshold % for uniq, second best must be at most this
fraction of best (default 0.96)
--commonGeneFile s
file name in which to write cases where one read maps
several different genes
--pairBedFile s
file name of pairedness coverage: a BED format file in
which for each position the number of filtered read pairs is reported that
contain the position in or between the reads
--pairwiseAlignments
use in case alignments were done in pairwise fashion
(default: 0)
AUGUSTUS was written by M. Stanke, O. Keller, S. König, L.
Gerischer and L. Romoth.
An exhaustive documentation can be found in the file
/usr/share/augustus/README.TXT.