|
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: Splits sequences according to their taxonomy,
|
|
jpayne@69
|
9 as determined by the sequence name. Sequences should
|
|
jpayne@69
|
10 be labeled with a gi number, NCBI taxID, or species name.
|
|
jpayne@69
|
11
|
|
jpayne@69
|
12 Usage: splitbytaxa.sh in=<input file> out=<output pattern> tree=<tree file> table=<table file> level=<name or number>
|
|
jpayne@69
|
13
|
|
jpayne@69
|
14 Input may be fasta or fastq, compressed or uncompressed.
|
|
jpayne@69
|
15
|
|
jpayne@69
|
16
|
|
jpayne@69
|
17 Standard parameters:
|
|
jpayne@69
|
18 in=<file> Primary input.
|
|
jpayne@69
|
19 out=<file> Output pattern; must contain % symbol.
|
|
jpayne@69
|
20 overwrite=f (ow) Set to false to force the program to abort rather than
|
|
jpayne@69
|
21 overwrite an existing file.
|
|
jpayne@69
|
22 showspeed=t (ss) Set to 'f' to suppress display of processing speed.
|
|
jpayne@69
|
23 ziplevel=2 (zl) Set to 1 (lowest) through 9 (max) to change compression
|
|
jpayne@69
|
24 level; lower compression is faster.
|
|
jpayne@69
|
25
|
|
jpayne@69
|
26 Processing parameters:
|
|
jpayne@69
|
27 level=phylum Taxonomic level, such as phylum. Filtering will operate on
|
|
jpayne@69
|
28 sequences within the same taxonomic level as specified ids.
|
|
jpayne@69
|
29 tree= A taxonomic tree made by TaxTree, such as tree.taxtree.gz.
|
|
jpayne@69
|
30 table= A table translating gi numbers to NCBI taxIDs.
|
|
jpayne@69
|
31 Only needed if gi numbers will be used.
|
|
jpayne@69
|
32 On Genepool, use 'tree=auto table=auto'.
|
|
jpayne@69
|
33 * Note *
|
|
jpayne@69
|
34 Tree and table files are in /global/projectb/sandbox/gaag/bbtools/tax
|
|
jpayne@69
|
35 For non-Genepool users, or to make new ones, use taxtree.sh and gitable.sh
|
|
jpayne@69
|
36
|
|
jpayne@69
|
37 Java Parameters:
|
|
jpayne@69
|
38 -Xmx This will set Java's memory usage, overriding automatic
|
|
jpayne@69
|
39 memory detection.
|
|
jpayne@69
|
40 -Xmx20g will specify 20 gigs of RAM, and -Xmx200m will specify
|
|
jpayne@69
|
41 200 megs. The max is typically 85% of physical memory.
|
|
jpayne@69
|
42 -eoom This flag will cause the process to exit if an out-of-memory
|
|
jpayne@69
|
43 exception occurs. Requires Java 8u92+.
|
|
jpayne@69
|
44 -da Disable assertions.
|
|
jpayne@69
|
45
|
|
jpayne@69
|
46 Please contact Brian Bushnell at bbushnell@lbl.gov if you encounter any problems.
|
|
jpayne@69
|
47 "
|
|
jpayne@69
|
48 }
|
|
jpayne@69
|
49
|
|
jpayne@69
|
50 #This block allows symlinked shellscripts to correctly set classpath.
|
|
jpayne@69
|
51 pushd . > /dev/null
|
|
jpayne@69
|
52 DIR="${BASH_SOURCE[0]}"
|
|
jpayne@69
|
53 while [ -h "$DIR" ]; do
|
|
jpayne@69
|
54 cd "$(dirname "$DIR")"
|
|
jpayne@69
|
55 DIR="$(readlink "$(basename "$DIR")")"
|
|
jpayne@69
|
56 done
|
|
jpayne@69
|
57 cd "$(dirname "$DIR")"
|
|
jpayne@69
|
58 DIR="$(pwd)/"
|
|
jpayne@69
|
59 popd > /dev/null
|
|
jpayne@69
|
60
|
|
jpayne@69
|
61 #DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )/"
|
|
jpayne@69
|
62 CP="$DIR""current/"
|
|
jpayne@69
|
63
|
|
jpayne@69
|
64 z="-Xmx4g"
|
|
jpayne@69
|
65 z2="-Xms4g"
|
|
jpayne@69
|
66 set=0
|
|
jpayne@69
|
67
|
|
jpayne@69
|
68 if [ -z "$1" ] || [[ $1 == -h ]] || [[ $1 == --help ]]; then
|
|
jpayne@69
|
69 usage
|
|
jpayne@69
|
70 exit
|
|
jpayne@69
|
71 fi
|
|
jpayne@69
|
72
|
|
jpayne@69
|
73 calcXmx () {
|
|
jpayne@69
|
74 source "$DIR""/calcmem.sh"
|
|
jpayne@69
|
75 setEnvironment
|
|
jpayne@69
|
76 parseXmx "$@"
|
|
jpayne@69
|
77 if [[ $set == 1 ]]; then
|
|
jpayne@69
|
78 return
|
|
jpayne@69
|
79 fi
|
|
jpayne@69
|
80 freeRam 1000m 84
|
|
jpayne@69
|
81 z="-Xmx${RAM}m"
|
|
jpayne@69
|
82 z2="-Xms${RAM}m"
|
|
jpayne@69
|
83 }
|
|
jpayne@69
|
84 calcXmx "$@"
|
|
jpayne@69
|
85
|
|
jpayne@69
|
86 splitbytaxa() {
|
|
jpayne@69
|
87 local CMD="java $EA $EOOM $z -cp $CP tax.SplitByTaxa $@"
|
|
jpayne@69
|
88 echo $CMD >&2
|
|
jpayne@69
|
89 eval $CMD
|
|
jpayne@69
|
90 }
|
|
jpayne@69
|
91
|
|
jpayne@69
|
92 splitbytaxa "$@"
|
