|
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 August 6, 2019
|
|
jpayne@69
|
7
|
|
jpayne@69
|
8 Description: Creates a mutant version of a genome.
|
|
jpayne@69
|
9
|
|
jpayne@69
|
10 Usage: mutate.sh in=<input file> out=<output file> id=<identity>
|
|
jpayne@69
|
11
|
|
jpayne@69
|
12 I/O parameters:
|
|
jpayne@69
|
13 in=<file> Input genome.
|
|
jpayne@69
|
14 out=<file> Output mutant genome.
|
|
jpayne@69
|
15 vcf=<file> Output VCF file showing variations added.
|
|
jpayne@69
|
16 overwrite=f (ow) Set to false to force the program to abort rather than
|
|
jpayne@69
|
17 overwrite an existing file.
|
|
jpayne@69
|
18 ziplevel=2 (zl) Set to 1 (lowest) through 9 (max) to change compression
|
|
jpayne@69
|
19 level; lower compression is faster.
|
|
jpayne@69
|
20
|
|
jpayne@69
|
21 Processing parameters:
|
|
jpayne@69
|
22 subrate=0 Substitution rate, 0 to 1.
|
|
jpayne@69
|
23 indelrate=0 Indel rate, 0 to 1.
|
|
jpayne@69
|
24 maxindel=1 Max indel length.
|
|
jpayne@69
|
25 indelspacing=10 Minimum distance between subsequent indels.
|
|
jpayne@69
|
26 id=1 Target identity, 0 to 1; 1 means 100%.
|
|
jpayne@69
|
27 If this is used it will override subrate and indelrate;
|
|
jpayne@69
|
28 99% of the mutations will be substitutions, and 1% indels.
|
|
jpayne@69
|
29 fraction=1 Genome fraction, 0 to 1; 1 means 100%. A lower fraction
|
|
jpayne@69
|
30 will randomly select that fraction on a per-sequence basis,
|
|
jpayne@69
|
31 possibly incurring one chimeric junction per sequence.
|
|
jpayne@69
|
32 Not compatible with VCF output.
|
|
jpayne@69
|
33 period=-1 If positive, place exactly one mutation every X bases.
|
|
jpayne@69
|
34 prefix= Set this flag to rename the new contigs with this prefix
|
|
jpayne@69
|
35 and a number.
|
|
jpayne@69
|
36 amino=f Treat the input as amino acid sequence.
|
|
jpayne@69
|
37 ploidy=1 Set the ploidy. ploidy>1 allows heterozygous mutations.
|
|
jpayne@69
|
38 This will create one copy of each input sequence per ploidy.
|
|
jpayne@69
|
39 hetrate=0.5 If polyploid, fraction of mutations that are heterozygous.
|
|
jpayne@69
|
40 nohomopolymers=f If true, prevent indels in homopolymers that lead to
|
|
jpayne@69
|
41 ambiguous variant calls. For example, inserting A between
|
|
jpayne@69
|
42 AC or deleting T from TTTT. This is mainly for grading
|
|
jpayne@69
|
43 purposes. It does not fully solve the problem, but greatly
|
|
jpayne@69
|
44 improves concordance (reducing disagreements by 70%).
|
|
jpayne@69
|
45
|
|
jpayne@69
|
46 Java Parameters:
|
|
jpayne@69
|
47 -Xmx This will set Java's memory usage, overriding autodetection.
|
|
jpayne@69
|
48 -Xmx20g will specify 20 gigs of RAM, and -Xmx200m will
|
|
jpayne@69
|
49 specify 200 megs. The max is typically 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 mutate() {
|
|
jpayne@69
|
95 local CMD="java $EA $EOOM $z $z2 -cp $CP jgi.MutateGenome $@"
|
|
jpayne@69
|
96 echo $CMD >&2
|
|
jpayne@69
|
97 eval $CMD
|
|
jpayne@69
|
98 }
|
|
jpayne@69
|
99
|
|
jpayne@69
|
100 mutate "$@"
|
