-
Notifications
You must be signed in to change notification settings - Fork 34
/
Copy pathVarScan.v2.4.4.description.txt
122 lines (97 loc) · 6.25 KB
/
VarScan.v2.4.4.description.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
RELEASE NOTES FOR VARSCAN V2.4.4
25-Jul-2019
VarScan v2.4.4 is the current release
VarScan v2.4.0 was the first release to VarScan's new home at GitHub, http://dkoboldt.github.io/varscan/
VarScan v2.3.9 and prior releases will persist on SourceForge, as will the support forums and other resources.
LICENSE
VarScan 2 is free for non-commercial use by academic, government, and non-profit/not-for-profit institutions.
A commercial version of the software is available, and licensed through the Office of Technology Management at
Washington University School of Medicine. For more information please visit their website at:
https://otm.wustl.edu/for-industry/tools/
or contact them by e-mail:
otm@wustl.edu
VERSION 2.4.4 CHANGES
This release includes minor bugfixes for user-reported issues and some additional parameters:
1.) The --max-somatic-p parameter was not being used in fpfilter, so setting it had no effect. This has been corrected. The default value
was changed to 1.0 so that this parameter is not applied by default. Thus, existing pipelines should not produce different results unless
a value is specified. Note that the --max-somatic-p-depth parameter [default: 10] specifies the minimum total read depth in tumor required
for the somatic P-value to be evaluated. This ensures that good calls at low coverage won't be discarded on the basis of p-value.
2.) Added a --min-mmqs-diff parameter to fpfilter, so that users can remove false positives with low/negative MMQSdiff values.
3.) Fixed processSomatic for VCF to infer read counts from RD/AD fields
4.) In somatic mode, the normal is now only called for the variant allele called in the tumor consensus.
5.) Made a change to indel type representation in VarScan.java that had caused some VCF formatting issues.
REMINDER: PLEASE USE THE FALSE POSITIVE FILTER (fpfilter)
We recommend processing initial VarScan output with processSomatic, and then applying fpfilter to high-confidence Somatic calls.
Although the "somaticFilter" function will continue to be available, we recommend using the above approach instead.
The scientific basis of this filter is described in the VarScan 2 publication. It will improve
the precision of variant and mutation calling by removing artifacts associated with short-read alignment.
-For somatic mutations, generate bam-readcounts with the Tumor BAM. For LOH and Germline, generate readcounts with the Normal BAM
-For de novo mutations (trio calling), generate readcounts with the child BAM.
The filter requires the bam-readcount utility: https://github.com/genome/bam-readcount
USAGE: java -jar VarScan.jar fpfilter [variant file] [readcount file] OPTIONS
variant file - A file of SNPs or indels in VarScan-native or VCF format
readcount file - The output file from bam-readcount for those positions
***For detailed filtering instructions, please visit http://varscan.sourceforge.net***
OPTIONS:
--output-file Optional output file for filter-pass variants
--filtered-file Optional output file for filter-fail variants
--dream3-settings If set to 1, optimizes filter parameters based on TCGA-ICGC DREAM-3 SNV Challenge results
--keep-failures If set to 1, includes failures in the output file
FILTERING PARAMETERS:
--min-var-count Minimum number of variant-supporting reads [4]
--min-var-count-lc Minimum number of variant-supporting reads when depth below somaticPdepth [2]
--min-var-freq Minimum variant allele frequency [0.05]
--max-somatic-p Maximum somatic p-value [1.0]
--max-somatic-p-depth Depth required to test max somatic p-value [10]
--min-ref-readpos Minimum average read position of ref-supporting reads [0.1]
--min-var-readpos Minimum average read position of var-supporting reads [0.1]
--min-ref-dist3 Minimum average distance to effective 3' end (ref) [0.1]
--min-var-dist3 Minimum average distance to effective 3' end (var) [0.1]
--min-strandedness Minimum fraction of variant reads from each strand [0.01]
--min-strand-reads Minimum allele depth required to perform the strand tests [5]
--min-ref-basequal Minimum average base quality for ref allele [15]
--min-var-basequal Minimum average base quality for var allele [15]
--min-ref-avgrl Minimum average trimmed read length for ref allele [90]
--min-var-avgrl Minimum average trimmed read length for var allele [90]
--max-rl-diff Maximum average relative read length difference (ref - var) [0.25]
--max-ref-mmqs Maximum mismatch quality sum of reference-supporting reads [100]
--max-var-mmqs Maximum mismatch quality sum of variant-supporting reads [100]
--max-mmqs-diff Maximum average mismatch quality sum (var - ref) [50]
--min-ref-mapqual Minimum average mapping quality for ref allele [15]
--min-var-mapqual Minimum average mapping quality for var allele [15]
--max-mapqual-diff Maximum average mapping quality (ref - var) [50]
DREAM-3 SETTINGS FOR FPFILTER
Please note the --dream3-settings parameter for fpfilter, which (if set to 1) will optimize the
false positive filter settings based on the fine-tuning we did for the TCGA-ICGC DREAM-3
SNV Challenge. See the "in silico 3" dataset described here:
https://www.synapse.org/#!Synapse:syn312572/wiki/62018
This dataset modeled 100% tumor purity, but three subclones at 50%, 33%, and 20% variant allele frequency.
Optimal VarScan settings were established as follows:
For SAMtools:
mpileup -B (disables BAQ)
For VarScan somatic:
--min-coverage 3 --min-var-freq 0.08 --p-value 0.10 --somatic-p-value 0.05 --strand-filter 0
For VarScan fpfilter:
--min-var-count = 3
--min-var-count-lc = 1
--min-strandedness = 0
--min-var-basequal = 30
--min-ref-readpos = 0.20
--min-ref-dist3 = 0.20
--min-var-readpos = 0.15
--min-var-dist3 = 0.15
--max-rl-diff = 0.05
--max-mapqual-diff = 10
--min-ref-mapqual = 20
--min-var-mapqual = 30
--max-var-mmqs = 100
--max-ref-mmqs = 50
CITING VARSCAN
If you use VarScan, please note the version number and cite this publication along with the
version-appropriate URL:
Koboldt DC, Zhang Q, Larson DE, Shen D, McLellan MD, Lin L, Miller CA, Mardis ER, Ding L, Wilson RK.
VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing.
Genome Res. 2012 Mar;22(3):568-76. doi: 10.1101/gr.129684.111.
https://github.com/dkoboldt/varscan (v2.4.0 and beyond)
or
http://varscan.sourceforge.net (v2.3.9 and before)