forked from dkoboldt/varscan
-
Notifications
You must be signed in to change notification settings - Fork 0
/
VarScan.v2.4.3.description.txt
117 lines (91 loc) · 6.02 KB
/
VarScan.v2.4.3.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
RELEASE NOTES FOR VARSCAN V2.4.3
01-Dec-2016
VarScan v2.4.3 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 [email protected].
VERSION 2.4.3 CHANGES
The principal change to this version changes how normal read counts are compiled during somatic mutation calling.
The remaining changes are minor bugfixes for user-reported issues:
1.) Some warnings printed by default to STDOUT (such as "Resetting normal file") are now only reported if --verbose 1 is set.
2.) Added a --min-mmqs-diff parameter to fpfilter, so that users can remove false positives with low/negative MMQSdiff values.
3.) Corrected some archaic code that represented indels like INS-3-ACG in VarScan native output. Native output uses more common
formats (+ACG) to represent the variant allele.
4.) Fixed the parsing exception errors in VarScan copynumber due to 8-column mpileup files (a correct line has 9 columns).
5.) Added a new option, --max-basequal-diff, to the fpfilter command (default value: 50).
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 [0.05]
--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)