|
jpayne@69
|
1 #!/bin/bash
|
|
jpayne@69
|
2
|
|
jpayne@69
|
3 usage(){
|
|
jpayne@69
|
4 echo "
|
|
jpayne@69
|
5 Written by Brian Bushnell
|
|
jpayne@69
|
6 Last modified Jan 7, 2020
|
|
jpayne@69
|
7
|
|
jpayne@69
|
8 Description: Prints the full taxonomy of a string.
|
|
jpayne@69
|
9 String may be a gi number, NCBI taxID, or Latin name.
|
|
jpayne@69
|
10 An NCBI identifier should just be a number or ncbi|number.
|
|
jpayne@69
|
11 A gi number should be gi|number.
|
|
jpayne@69
|
12 Please read bbmap/docs/guides/TaxonomyGuide.txt for more information.
|
|
jpayne@69
|
13 Not: It is more convenient to use taxonomy.jgi-psf.org.
|
|
jpayne@69
|
14
|
|
jpayne@69
|
15 Usage: taxonomy.sh tree=<tree file> <identifier>
|
|
jpayne@69
|
16 Alternate usage: taxonomy.sh tree=<tree file> in=<file>
|
|
jpayne@69
|
17
|
|
jpayne@69
|
18 Usage examples:
|
|
jpayne@69
|
19 taxonomy.sh tree=tree.taxtree.gz homo_sapiens canis_lupus 9606
|
|
jpayne@69
|
20 taxonomy.sh tree=tree.taxtree.gz gi=gitable.int1.d.gz in=refseq.fasta
|
|
jpayne@69
|
21
|
|
jpayne@69
|
22 Processing parameters:
|
|
jpayne@69
|
23 in=<file> A file containing named sequences, or just the names.
|
|
jpayne@69
|
24 out=<file> Output file. If blank, use stdout.
|
|
jpayne@69
|
25 tree=<file> Specify a TaxTree file like tree.taxtree.gz.
|
|
jpayne@69
|
26 On Genepool, use 'auto'.
|
|
jpayne@69
|
27 gi=<file> Specify a gitable file like gitable.int1d.gz. Only needed
|
|
jpayne@69
|
28 if gi numbers will be used. On Genepool, use 'auto'.
|
|
jpayne@69
|
29 accession= Specify one or more comma-delimited NCBI accession to taxid
|
|
jpayne@69
|
30 files. Only needed if accesions will be used; requires ~45GB
|
|
jpayne@69
|
31 of memory. On Genepool, use 'auto'.
|
|
jpayne@69
|
32 level=null Set to a taxonomic level like phylum to just print that level.
|
|
jpayne@69
|
33 minlevel=-1 For multi-level printing, do not print levels below this.
|
|
jpayne@69
|
34 maxlevel=life For multi-level printing, do not print levels above this.
|
|
jpayne@69
|
35 silva=f Parse headers using Silva or semicolon-delimited syntax.
|
|
jpayne@69
|
36 taxpath=auto Set the path to taxonomy files; auto only works at NERSC.
|
|
jpayne@69
|
37
|
|
jpayne@69
|
38 Parameters without an '=' symbol will be considered organism identifiers.
|
|
jpayne@69
|
39
|
|
jpayne@69
|
40 * Note *
|
|
jpayne@69
|
41 Tree and table files are in /global/projectb/sandbox/gaag/bbtools/tax
|
|
jpayne@69
|
42 For non-Genepool users, or to make new ones, use taxtree.sh and gitable.sh
|
|
jpayne@69
|
43
|
|
jpayne@69
|
44 Java Parameters:
|
|
jpayne@69
|
45 -Xmx This will set Java's memory usage,
|
|
jpayne@69
|
46 overriding autodetection.
|
|
jpayne@69
|
47 -Xmx20g will specify 20 gigs of RAM, and -Xmx200m will specify
|
|
jpayne@69
|
48 200 megs. The max is typically 85% of physical memory.
|
|
jpayne@69
|
49 -eoom This flag will cause the process to exit if an out-of-memory
|
|
jpayne@69
|
50 exception occurs. Requires Java 8u92+.
|
|
jpayne@69
|
51 -da Disable assertions.
|
|
jpayne@69
|
52
|
|
jpayne@69
|
53 Please contact Brian Bushnell at bbushnell@lbl.gov if you encounter any problems.
|
|
jpayne@69
|
54 "
|
|
jpayne@69
|
55 }
|
|
jpayne@69
|
56
|
|
jpayne@69
|
57 #This block allows symlinked shellscripts to correctly set classpath.
|
|
jpayne@69
|
58 pushd . > /dev/null
|
|
jpayne@69
|
59 DIR="${BASH_SOURCE[0]}"
|
|
jpayne@69
|
60 while [ -h "$DIR" ]; do
|
|
jpayne@69
|
61 cd "$(dirname "$DIR")"
|
|
jpayne@69
|
62 DIR="$(readlink "$(basename "$DIR")")"
|
|
jpayne@69
|
63 done
|
|
jpayne@69
|
64 cd "$(dirname "$DIR")"
|
|
jpayne@69
|
65 DIR="$(pwd)/"
|
|
jpayne@69
|
66 popd > /dev/null
|
|
jpayne@69
|
67
|
|
jpayne@69
|
68 #DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )/"
|
|
jpayne@69
|
69 CP="$DIR""current/"
|
|
jpayne@69
|
70
|
|
jpayne@69
|
71 z="-Xmx8g"
|
|
jpayne@69
|
72 z2="-Xms8g"
|
|
jpayne@69
|
73 set=0
|
|
jpayne@69
|
74
|
|
jpayne@69
|
75 if [ -z "$1" ] || [[ $1 == -h ]] || [[ $1 == --help ]]; then
|
|
jpayne@69
|
76 usage
|
|
jpayne@69
|
77 exit
|
|
jpayne@69
|
78 fi
|
|
jpayne@69
|
79
|
|
jpayne@69
|
80 calcXmx () {
|
|
jpayne@69
|
81 source "$DIR""/calcmem.sh"
|
|
jpayne@69
|
82 setEnvironment
|
|
jpayne@69
|
83 parseXmx "$@"
|
|
jpayne@69
|
84 if [[ $set == 1 ]]; then
|
|
jpayne@69
|
85 return
|
|
jpayne@69
|
86 fi
|
|
jpayne@69
|
87 freeRam 2000m 84
|
|
jpayne@69
|
88 z="-Xmx${RAM}m"
|
|
jpayne@69
|
89 z2="-Xms${RAM}m"
|
|
jpayne@69
|
90 }
|
|
jpayne@69
|
91 calcXmx "$@"
|
|
jpayne@69
|
92
|
|
jpayne@69
|
93 taxonomy() {
|
|
jpayne@69
|
94 local CMD="java $EA $EOOM $z -cp $CP tax.PrintTaxonomy $@"
|
|
jpayne@69
|
95 echo $CMD >&2
|
|
jpayne@69
|
96 eval $CMD
|
|
jpayne@69
|
97 }
|
|
jpayne@69
|
98
|
|
jpayne@69
|
99 taxonomy "$@"
|
