|
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 January 21, 2020
|
|
jpayne@69
|
7
|
|
jpayne@69
|
8 Description: Generates synthetic PacBio reads to mimic the chimeric
|
|
jpayne@69
|
9 inverted repeats from 'triangle reads', aka 'ice cream cones' -
|
|
jpayne@69
|
10 reads missing one adapter.
|
|
jpayne@69
|
11
|
|
jpayne@69
|
12 Usage: icecreammaker.sh in=<file> out=<file> reads=100k minlen=500 maxlen=5000
|
|
jpayne@69
|
13
|
|
jpayne@69
|
14 Standard parameters:
|
|
jpayne@69
|
15 in=<file> A reference genome fasta (optional).
|
|
jpayne@69
|
16 out=<file> Synthetic read output.
|
|
jpayne@69
|
17 idhist=<file> Identity histogram output.
|
|
jpayne@69
|
18 overwrite=f (ow) Set to false to force the program to abort rather than
|
|
jpayne@69
|
19 overwrite an existing file.
|
|
jpayne@69
|
20 ziplevel=2 (zl) Set to 1 (lowest) through 9 (max) to change compression
|
|
jpayne@69
|
21 level; lower compression is faster.
|
|
jpayne@69
|
22
|
|
jpayne@69
|
23 Length parameters:
|
|
jpayne@69
|
24 NOTE: "length" parameters dictate subread length (for normal reads).
|
|
jpayne@69
|
25 "Movie" parameters dictate sequence length (with concatenated subreads).
|
|
jpayne@69
|
26
|
|
jpayne@69
|
27 minlen=500 (minlength) Minimum length of genomic sequence molecules.
|
|
jpayne@69
|
28 maxlen=5000 (maxlength) Maximum length of genomic sequence molecules.
|
|
jpayne@69
|
29 len= (length) Set minlen and maxlen to the same number.
|
|
jpayne@69
|
30 minmovie=500 (minmov) Minimum length of movies.
|
|
jpayne@69
|
31 maxmovie=40k (maxmov) Maximum length of movies.
|
|
jpayne@69
|
32 movie= (mov) Set minmov and maxmov to the same number.
|
|
jpayne@69
|
33
|
|
jpayne@69
|
34 Ice cream parameters:
|
|
jpayne@69
|
35 missingrate=0 (missing) Fraction of reads missing an adapter.
|
|
jpayne@69
|
36 hiddenrate=0 (hidden) Fraction of adapters not detected.
|
|
jpayne@69
|
37 bothends=f Allow missing or hiddden adapters on both ends.
|
|
jpayne@69
|
38
|
|
jpayne@69
|
39 Other parameters:
|
|
jpayne@69
|
40 zmws (reads) Number of ZMWs to generate. There are actually
|
|
jpayne@69
|
41 multiple subreads per zmw.
|
|
jpayne@69
|
42 ccs=f Make CCS reads (one read per ZMW, full pass only).
|
|
jpayne@69
|
43 You still need to specify the error rate.
|
|
jpayne@69
|
44 gc=0.6 If a random genome is generated, use this GC fraction.
|
|
jpayne@69
|
45 genomesize=10m If a random genome is generated, make it this big.
|
|
jpayne@69
|
46 irrate=0.0 Add inverted repeats until this fraction of the genome
|
|
jpayne@69
|
47 is inverted repeats.
|
|
jpayne@69
|
48 irminlen=500 Minimum length of inverted repeats.
|
|
jpayne@69
|
49 irmaxlen=5000 Maximum length of inverted repeats
|
|
jpayne@69
|
50 irlen= Set minirlen and maxirlen to the same number.
|
|
jpayne@69
|
51 miner=0.05 (minerrorrate) Minimum error rate.
|
|
jpayne@69
|
52 maxer=0.28 (maxerrorrate) Maximum error rate.
|
|
jpayne@69
|
53 er= (errorrate) Set minerrorrate and maxerrorrate.
|
|
jpayne@69
|
54 NOTE: You can alternatively set minid, maxid, or id.
|
|
jpayne@69
|
55
|
|
jpayne@69
|
56
|
|
jpayne@69
|
57 Java Parameters:
|
|
jpayne@69
|
58 -Xmx This will set Java's memory usage, overriding autodetection.
|
|
jpayne@69
|
59 -Xmx20g will specify 20 gigs of RAM, and -Xmx200m will
|
|
jpayne@69
|
60 specify 200 megs. The max is typically 85% of physical memory.
|
|
jpayne@69
|
61 -eoom This flag will cause the process to exit if an out-of-memory
|
|
jpayne@69
|
62 exception occurs. Requires Java 8u92+.
|
|
jpayne@69
|
63 -da Disable assertions.
|
|
jpayne@69
|
64
|
|
jpayne@69
|
65 Please contact Brian Bushnell at bbushnell@lbl.gov if you encounter any problems.
|
|
jpayne@69
|
66 "
|
|
jpayne@69
|
67 }
|
|
jpayne@69
|
68
|
|
jpayne@69
|
69 #This block allows symlinked shellscripts to correctly set classpath.
|
|
jpayne@69
|
70 pushd . > /dev/null
|
|
jpayne@69
|
71 DIR="${BASH_SOURCE[0]}"
|
|
jpayne@69
|
72 while [ -h "$DIR" ]; do
|
|
jpayne@69
|
73 cd "$(dirname "$DIR")"
|
|
jpayne@69
|
74 DIR="$(readlink "$(basename "$DIR")")"
|
|
jpayne@69
|
75 done
|
|
jpayne@69
|
76 cd "$(dirname "$DIR")"
|
|
jpayne@69
|
77 DIR="$(pwd)/"
|
|
jpayne@69
|
78 popd > /dev/null
|
|
jpayne@69
|
79
|
|
jpayne@69
|
80 #DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )/"
|
|
jpayne@69
|
81 CP="$DIR""current/"
|
|
jpayne@69
|
82
|
|
jpayne@69
|
83 z="-Xmx2g"
|
|
jpayne@69
|
84 z2="-Xms2g"
|
|
jpayne@69
|
85 set=0
|
|
jpayne@69
|
86
|
|
jpayne@69
|
87 if [ -z "$1" ] || [[ $1 == -h ]] || [[ $1 == --help ]]; then
|
|
jpayne@69
|
88 usage
|
|
jpayne@69
|
89 exit
|
|
jpayne@69
|
90 fi
|
|
jpayne@69
|
91
|
|
jpayne@69
|
92 calcXmx () {
|
|
jpayne@69
|
93 source "$DIR""/calcmem.sh"
|
|
jpayne@69
|
94 setEnvironment
|
|
jpayne@69
|
95 parseXmx "$@"
|
|
jpayne@69
|
96 if [[ $set == 1 ]]; then
|
|
jpayne@69
|
97 return
|
|
jpayne@69
|
98 fi
|
|
jpayne@69
|
99 freeRam 2000m 42
|
|
jpayne@69
|
100 z="-Xmx${RAM}m"
|
|
jpayne@69
|
101 z2="-Xms${RAM}m"
|
|
jpayne@69
|
102 }
|
|
jpayne@69
|
103 calcXmx "$@"
|
|
jpayne@69
|
104
|
|
jpayne@69
|
105 icecreammaker() {
|
|
jpayne@69
|
106 local CMD="java $EA $EOOM $z -cp $CP icecream.IceCreamMaker $@"
|
|
jpayne@69
|
107 echo $CMD >&2
|
|
jpayne@69
|
108 eval $CMD
|
|
jpayne@69
|
109 }
|
|
jpayne@69
|
110
|
|
jpayne@69
|
111 icecreammaker "$@"
|
