Mercurial > repos > ryanmorin > nextgen_variant_identification
comparison SNV/SNVMix2_source/SNVMix2-v0.12.1-rc1/README @ 0:74f5ea818cea
Uploaded
| author | ryanmorin |
|---|---|
| date | Wed, 12 Oct 2011 19:50:38 -0400 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| -1:000000000000 | 0:74f5ea818cea |
|---|---|
| 1 *** SNVMix2 *** | |
| 2 | |
| 3 This version of SNVMix2 has has been tested under Linux and Mac OS X. | |
| 4 | |
| 5 To build: | |
| 6 | |
| 7 > unzip -x ../SNVMix2-v{VERSION}.zip | |
| 8 > cd SNVmix2-v{VERSION}/ | |
| 9 > make | |
| 10 | |
| 11 Binary file is called "SNVMix2", for help run with -h flag | |
| 12 | |
| 13 > ./SNVMix2 -h | |
| 14 | |
| 15 You can copy that file to a preferred location. | |
| 16 | |
| 17 SNVMix2 defaults to reading from standard input if flag '-i' is not specified, | |
| 18 so you can use it in pipe mode next to a MAQ or samtools pileup command, saving | |
| 19 storage space. Same applies for standard output and the "-o" flag while Classifying. | |
| 20 | |
| 21 Parameter '-m <file>' is used to read the model parameters when Classifying (-C) | |
| 22 or write them in Train mode (-T) | |
| 23 | |
| 24 Different models for base and mapping qualities are available using the "-t" flag, | |
| 25 SNVMix1 mode can be accessed by selecting "-t SNVMix1" | |
| 26 | |
| 27 Pileup file should be generated with base and mapping qualities but without consensus, | |
| 28 such as the one obtained when running, with samtools-0.1.4: | |
| 29 samtools pileup -s -l <in_list.txt> -f <ref.fa> <in.bam> | |
| 30 | |
| 31 or with maq-0.7.0: | |
| 32 maq pileup -v <ref.bfa> <in.map> | |
| 33 | |
| 34 | |
| 35 A perl script is provided to filter SNVMix2 result according to a specified | |
| 36 threshold, help on this script can be found running: | |
| 37 | |
| 38 > ./misc/snvmix2summary.pl -h | |
| 39 | |
| 40 This script can filter candidate SNVs using two methods specified by the '-c <TYPE>' | |
| 41 flag: | |
| 42 | |
| 43 '-c 2' Will consider only two classes, either homozygote for the reference allele | |
| 44 ( p(AA) ) or not ( p(AB) + p(BB) ) | |
| 45 | |
| 46 '-c 3' Will consider the three clases p(AA), p(AB), p(BB) separately | |
| 47 | |
| 48 | |
| 49 The output still still retains the unmodified probability values, so p_AB can still | |
| 50 be distinguished from p_BB in case '-c 2' is used. | |
| 51 | |
| 52 | |
| 53 | |
| 54 | |
| 55 | |
| 56 Notes: | |
| 57 | |
| 58 A working gcc compiler is needed and under Linux libc >= 4.6.27 is required. | |
| 59 | |
| 60 Be careful if you copy & paste pileup files, as some text editors tend to change | |
| 61 "tabs" for spaces, which is needed as the field delimiter. This can cause problems | |
| 62 in the parser. | |
| 63 | |
| 64 // | |
| 65 Comments and Questions: | |
| 66 Sohrab Shah sshah@bccrc.ca | |
| 67 Rodrigo Goya rgoya@bcgsc.ca | |
| 68 | |
| 69 // | |
| 70 LICENSE: this software is distributed under the MIT license |