|
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 September 4, 2019
|
|
jpayne@69
|
7
|
|
jpayne@69
|
8 Description: Makes a representative set of taxa from all-to-all identity
|
|
jpayne@69
|
9 comparison. Input should be in 3+ column TSV format (first 3 are required):
|
|
jpayne@69
|
10 (query, ref, ANI, qsize, rsize, qbases, rbases)
|
|
jpayne@69
|
11 ...as produced by CompareSketch with format=3 and usetaxidname.
|
|
jpayne@69
|
12 Additional columns are allowed and will be ignored.
|
|
jpayne@69
|
13
|
|
jpayne@69
|
14 Usage: representative.sh in=<input file> out=<output file>
|
|
jpayne@69
|
15
|
|
jpayne@69
|
16 Parameters:
|
|
jpayne@69
|
17 overwrite=f (ow) Set to false to force the program to abort rather than
|
|
jpayne@69
|
18 overwrite an existing file.
|
|
jpayne@69
|
19 threshold=0 Ignore edges under threshold value. This also affects the
|
|
jpayne@69
|
20 choice of centroids; a high threshold gives more weight to
|
|
jpayne@69
|
21 higher-value edges.
|
|
jpayne@69
|
22 minratio=0 Ignores edges with a ratio below this value.
|
|
jpayne@69
|
23 invertratio=f Invert the ratio when greater than 1.
|
|
jpayne@69
|
24 printheader=t Print a header line in the output.
|
|
jpayne@69
|
25 printsize=t Print the size of retained nodes.
|
|
jpayne@69
|
26 printclusters=t Print the nodes subsumed by each retained node.
|
|
jpayne@69
|
27 minsize=0 Ignore nodes under this size (in unique kmers).
|
|
jpayne@69
|
28 maxsize=0 If positive, ignore nodes over this size (unique kmers).
|
|
jpayne@69
|
29 minbases=0 Ignore nodes under this size (in total bases).
|
|
jpayne@69
|
30 maxbases=0 If positive, ignore nodes over this size (total bases).
|
|
jpayne@69
|
31
|
|
jpayne@69
|
32 Taxonomy parameters:
|
|
jpayne@69
|
33 level= Taxonomic level, such as phylum. Filtering will operate on
|
|
jpayne@69
|
34 sequences within the same taxonomic level as specified ids.
|
|
jpayne@69
|
35 If not set, only matches to a node or its descendants will
|
|
jpayne@69
|
36 be considered.
|
|
jpayne@69
|
37 ids= Comma-delimited list of NCBI numeric IDs. Can also be a
|
|
jpayne@69
|
38 file with one taxID per line.
|
|
jpayne@69
|
39 names= Alternately, a list of names (such as 'Homo sapiens').
|
|
jpayne@69
|
40 Note that spaces need special handling.
|
|
jpayne@69
|
41 include=f 'f' will discard filtered sequences, 't' will keep them.
|
|
jpayne@69
|
42 tree=<file> Specify a TaxTree file like tree.taxtree.gz.
|
|
jpayne@69
|
43 On Genepool, use 'auto'.
|
|
jpayne@69
|
44
|
|
jpayne@69
|
45 Java Parameters:
|
|
jpayne@69
|
46 -Xmx This will set Java's memory usage, overriding autodetection.
|
|
jpayne@69
|
47 -Xmx20g will
|
|
jpayne@69
|
48 specify 20 gigs of RAM, and -Xmx200m will specify 200 megs.
|
|
jpayne@69
|
49 The max is typically around 85% of physical memory.
|
|
jpayne@69
|
50 -eoom This flag will cause the process to exit if an out-of-memory
|
|
jpayne@69
|
51 exception occurs. Requires Java 8u92+.
|
|
jpayne@69
|
52 -da Disable assertions.
|
|
jpayne@69
|
53
|
|
jpayne@69
|
54 Please contact Brian Bushnell at bbushnell@lbl.gov if you encounter any problems.
|
|
jpayne@69
|
55 "
|
|
jpayne@69
|
56 }
|
|
jpayne@69
|
57
|
|
jpayne@69
|
58 #This block allows symlinked shellscripts to correctly set classpath.
|
|
jpayne@69
|
59 pushd . > /dev/null
|
|
jpayne@69
|
60 DIR="${BASH_SOURCE[0]}"
|
|
jpayne@69
|
61 while [ -h "$DIR" ]; do
|
|
jpayne@69
|
62 cd "$(dirname "$DIR")"
|
|
jpayne@69
|
63 DIR="$(readlink "$(basename "$DIR")")"
|
|
jpayne@69
|
64 done
|
|
jpayne@69
|
65 cd "$(dirname "$DIR")"
|
|
jpayne@69
|
66 DIR="$(pwd)/"
|
|
jpayne@69
|
67 popd > /dev/null
|
|
jpayne@69
|
68
|
|
jpayne@69
|
69 #DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )/"
|
|
jpayne@69
|
70 CP="$DIR""current/"
|
|
jpayne@69
|
71
|
|
jpayne@69
|
72 z="-Xmx4g"
|
|
jpayne@69
|
73 z2="-Xms4g"
|
|
jpayne@69
|
74 set=0
|
|
jpayne@69
|
75
|
|
jpayne@69
|
76 if [ -z "$1" ] || [[ $1 == -h ]] || [[ $1 == --help ]]; then
|
|
jpayne@69
|
77 usage
|
|
jpayne@69
|
78 exit
|
|
jpayne@69
|
79 fi
|
|
jpayne@69
|
80
|
|
jpayne@69
|
81 calcXmx () {
|
|
jpayne@69
|
82 source "$DIR""/calcmem.sh"
|
|
jpayne@69
|
83 setEnvironment
|
|
jpayne@69
|
84 parseXmx "$@"
|
|
jpayne@69
|
85 if [[ $set == 1 ]]; then
|
|
jpayne@69
|
86 return
|
|
jpayne@69
|
87 fi
|
|
jpayne@69
|
88 freeRam 4000m 84
|
|
jpayne@69
|
89 z="-Xmx${RAM}m"
|
|
jpayne@69
|
90 z2="-Xms${RAM}m"
|
|
jpayne@69
|
91 }
|
|
jpayne@69
|
92 calcXmx "$@"
|
|
jpayne@69
|
93
|
|
jpayne@69
|
94 a_sample_mt() {
|
|
jpayne@69
|
95 local CMD="java $EA $EOOM $z -cp $CP jgi.RepresentativeSet $@"
|
|
jpayne@69
|
96 echo $CMD >&2
|
|
jpayne@69
|
97 eval $CMD
|
|
jpayne@69
|
98 }
|
|
jpayne@69
|
99
|
|
jpayne@69
|
100 a_sample_mt "$@"
|
