yandell-lab / vvp-pub Goto Github PK

View Code? Open in Web Editor NEW

19.0 9.0 10.0 1.1 MB

VVP (VAAST Variant Prioritizer) rapidly prioritizes genetic variants

C 99.81% Makefile 0.19%

variant-analysis

vvp-pub's Introduction

VVP

Variant prioritization / burden test. Version 1.5

INSTALL

DEPENDENCIES

Gnu scientific library (https://www.gnu.org/software/gsl/)
openmp compatible version of gcc. If your compiler (clang) is not, you can remove the -fopenmp flag in the Makefile. Change the line that looks like: CFLAGS = -lz -lm -O3 -lgsl -lgslcblas -fopenmp #-Wall to CFLAGS = -lz -lm -O3 -lgsl -lgslcblas #-fopenmp #-Wall
zlib (https://zlib.net)
make

BUILD

In the VVP directory:

make

Make will build 2 executables: build_background and VVP

Note: This has been built and run on Mac laptops and Linux servers.

EXAMPLE RUNNING VVP

To see available parameters of the executables, run with the -h option.

Before running VVP, a background must be built. From the VVP directory:

cd example

../build_background -i 1KG_cftr_background.recode.vep.vcf.gz -o 1KG.build -b 2500 -v CSQ,4,6,1,15

The build_background step produces output to stdout for each of the variants in the background vcf file. It also creates several different output files including extensions .bin, .chr_offsets.txt, .dist. These files contained information used by VVP.

To run prioritize variants using VVP (in the example folder):

../VVP -i target_spiked_simple.vcf.gz -d 1KG.build -v CSQ,4,6,1,15 1> target.spiked.vvp.out

target_spiked.vvp.out contains the vvp output.

PREPARE VCF FILE FOR ANALYSIS

The VVP pipeline does not support mulitallelic lines, these must first be decomposed. We recommend using vt decompose to accomplish this task (http://genome.sph.umich.edu/wiki/Vt).

Mandatory preprocessing of a vcf file includes multiallelic decomposition and VEP annotation. It is important to decompose BEFORE annnotating because of potential annotation collisions. Our recommended steps are to use vt to decompose and normalize variants followed by VEP annotation. No special options in VEP are required for the variant annotation. Testing has been done with VEP v82.

VVP BACKGROUND

A prebuilt background based on gnomAD (http://gnomad.broadinstitute.org/) for use with VVP can be downloaded here (2.5GB): https://s3-us-west-2.amazonaws.com/gnomad-vvp-background/gnomad.062717.build.tar.gz

VVP OUTPUT

VVP outputs a tab delimited file with 31 columns. The columns are the following:

column name	description
chr	chromosome
start	variant start coord
ref	reference allele
var	variant allele
gene	gene id
transcript	transcript id
hemi_score	raw variant score for hemizygous genotype
hemi_vvp	vvp score for hemizygous genotype
nhemi	number of hemizygous indivduals
hemi_indvs	list of hemizygous individuals
hemi_nocall	number of hemizygous nocalls
het_score	raw variant score for heterozygous genotype
het_vvp	vvp score for heterozygous genotype
nhet	number of heterozygous individuals
het_indvs	list of heterozygous individuals
het_nocall	number of heterozygous nocalls
hom_score	raw variant score for homozygous genotype
hom_vvp	vvp score for homozygous genotype
nhom	number of homozygous individuals
hom_indvs	list of homozygous individuals
hom_nocall	number of homozygous nocalls
coding_ind	1 if variant is coding, 0 otherwise
indel_ind	1 if variant is an indel, 0 otherwise
aa_score	amino acid weight
n_bhemi	number of hemizygous background individuals
n_bhet	number of heterozygous background individuals
n_bhom	number of homozygous background individuals
n_bnocall	number of alleles nocalled in background
bit_offset	byte offset to background
vid	variant id
ll_weight	optional extra weight

vvp-pub's People

Contributors

Stargazers

Watchers

Forkers

inambioinfoku raonyguimaraes haoziyeung eulertx scalavision mingcheninsz dlhuang cgwyx genieus chakeraloui

vvp-pub's Issues

How to obtain CSQ

Now，I get CSQ through vep, like this:

But when i run the shell

../VVP -i gnomad_chr_1_1000.vcf.gz -d gnomad.062717.build -v CSQ,4,6,1,15 1> target.spiked.vvp.out

The output target.spiked.vvp.out has nothing:

Using build_background for gnomAD v3

Hi,

I am having trouble creating a background using gnomAD v3.

https://storage.googleapis.com/gnomad-public/release/3.0/vcf/genomes/gnomad.genomes.r3.0.sites.vcf.bgz

Command ran:

build_background -i gnomad.genomes.r3.0.sites.vcf.bgz -b 1 -v CSQ,4,6,1,15 -o test.out

I receive the error:

#chr    start   ref     var     transcript      hemi_score      het_score       hom_score       nhemi   nhet    nhom    hemi_nocall     het_nocall      hom_nocall      coding_ind
VCF FORMAT PROBLEM:      vcf line has fewer than 10 columns, will skip: chr1:10031, C, AC=0;AN=53780;AF=0.00000e+00;lcr;variant_type=snv;n_alt_alleles=1;ReadPosRankSum=-1.38000e+00;MQRankSum=-5.72000e-01;RAW_MQ=6.39630e+04;DP=26;MQ_DP=52;VarDP=26;MQ=3.50722e+01;QD=2.96154e+00;FS=5.09715e+00;SB=21,6,3,3;InbreedingCoeff=-1.72592e-05;AS_VQSLOD=-9.41040e+00;NEGATIVE_TRAIN_SITE;culprit=AS_MQ;SOR=9.60000e-02;AC_asj_female=0;AN_asj_female=776;AF_asj_female=0.00000e+00;nhomalt_asj_female=0;AC_eas_female=0;AN_eas_female=558;AF_eas_female=0.00000e+00;nhomalt_eas_female=0;AC_afr_male=0;AN_afr_male=6700;AF_afr_male=0.00000e+00;nhomalt_afr_male=0;AC_female=0;AN_female=27974;AF_female=0.00000e+00;nhomalt_female=0;AC_fin_male=0;AN_fin_male=3278;AF_fin_male=0.00000e+00;nhomalt_fin_male=0;AC_oth_female=0;AN_oth_female=430;AF_oth_female=0.00000e+00;nhomalt_oth_female=0;AC_ami=0;AN_ami=350;AF_ami=0.00000e+00;nhomalt_ami=0;AC_oth=0;AN_oth=802;AF_oth=0.00000e+00;nhomalt_oth=0;AC_male=0;AN_male=25806;AF_male=0.00000e+00;nhomalt_male=0;AC_ami_female=0;AN_ami_female=150;AF_ami_female=0.00000e+00;nhomalt_ami_female=0;AC_afr=0;AN_afr=14854;AF_afr=0.00000e+00;nhomalt_afr=0;AC_eas_male=0;AN_eas_male=612;AF_eas_male=0.00000e+00;nhomalt_eas_male=0;AC_sas=0;AN_sas=606;AF_sas=0.00000e+00;nhomalt_sas=0;AC_nfe_female=0;AN_nfe_female=14256;AF_nfe_female=0.00000e+00;nhomalt_nfe_female=0;AC_asj_male=0;AN_asj_male=720;AF_asj_male=0.00000e+00;nhomalt_asj_male=0;AC_raw=2;AN_raw=115882;AF_raw=1.72589e-05;nhomalt_raw=0;AC_oth_male=0;AN_oth_male=372;AF_oth_male=0.00000e+00;nhomalt_oth_male=0;AC_nfe_male=0;AN_nfe_male=10036;AF_nfe_male=0.00000e+00;nhomalt_nfe_male=0;AC_asj=0;AN_asj=1496;AF_asj=0.00000e+00;nhomalt_asj=0;AC_amr_male=0;AN_amr_male=3392;AF_amr_male=0.00000e+00;nhomalt_amr_male=0;nhomalt=0;AC_amr_female=0;AN_amr_female=2454;AF_amr_female=0.00000e+00;nhomalt_amr_female=0;AC_sas_female=0;AN_sas_female=110;AF_sas_female=0.00000e+00;nhomalt_sas_female=0;AC_fin=0;AN_fin=4364;AF_fin=0.00000e+00;nhomalt_fin=0;AC_afr_female=0;AN_afr_female=8154;AF_afr_female=0.00000e+00;nhomalt_afr_female=0;AC_sas_male=0;AN_sas_male=496;AF_sas_male=0.00000e+00;nhomalt_sas_male=0;AC_amr=0;AN_amr=5846;AF_amr=0.00000e+00;nhomalt_amr=0;AC_nfe=0;AN_nfe=24292;AF_nfe=0.00000e+00;nhomalt_nfe=0;AC_eas=0;AN_eas=1170;AF_eas=0.00000e+00;nhomalt_eas=0;AC_ami_male=0;AN_ami_male=200;AF_ami_male=0.00000e+00;nhomalt_ami_male=0;AC_fin_female=0;AN_fin_female=1086;AF_fin_female=0.00000e+00;nhomalt_fin_female=0;faf95_afr=0.00000e+00;faf99_afr=0.00000e+00;faf95_sas=0.00000e+00;faf99_sas=0.00000e+00;faf95_adj=0.00000e+00;faf99_adj=0.00000e+00;faf95_amr=0.00000e+00;faf99_amr=0.00000e+00;faf95_nfe=0.00000e+00;faf99_nfe=0.00000e+00;faf95_eas=0.00000e+00;faf99_eas=0.00000e+00;vep=C|upstream_gene_variant|MODIFIER|DDX11L1|ENSG00000223972|Transcript|ENST00000450305|transcribed_unprocessed_pseudogene|||,C|upstream_gene_variant|MODIFIER|DDX11L1|ENSG00000223972|Transcript|ENST00000456328|processed_transcript|||,C|downstream_gene_variant|MODIFIER|WASH7P|ENSG00000227232|Transcript|ENST00000488147|unprocessed_pseudogene|||
Segmentation fault (core dumped)

Is build background not compatible with this new gnomAD version?

how to use background based on gnomAD ?

how to use background based on gnomAD?

lock up if VVP is run concurrently

When two or more VVP processes are run concurrently they may all need to open the same background for memory mapping. For some reason the background is opened for read and write rather than for read only and this locks up the processes so everything just hangs.

I've identified the relevant bits of code here
https://github.com/Yandell-Lab/VVP-pub/blob/master/search_binary_bkgrnd.c#L20
https://github.com/Yandell-Lab/VVP-pub/blob/master/search_binary_bkgrnd.c#L24
and here
https://github.com/Yandell-Lab/VVP-pub/blob/master/search_binary_bkgrnd.c#L61
https://github.com/Yandell-Lab/VVP-pub/blob/master/search_binary_bkgrnd.c#L64

I've made a patch which seems to be working although I must admit I have not looked deep into the code.

20c20
<     int fdSrc = open(bin_file, O_RDONLY, 0);
---
>     int fdSrc = open(bin_file, O_RDWR, 0);
24c24
<     mm_bin = (unsigned char *)mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE | MAP_POPULATE, fdSrc, 0);
---
>     mm_bin = (unsigned char *)mmap(NULL, st.st_size, PROT_READ | PROT_WRITE, MAP_SHARED, fdSrc, 0);
61c61
<     int fdSrc = open(bit_file, O_RDONLY, 0);
---
>     int fdSrc = open(bit_file, O_RDWR, 0);
64c64
<     mm_bits = (unsigned char *)mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE | MAP_POPULATE, fdSrc, 0);
---
>     mm_bits = (unsigned char *)mmap(NULL, st.st_size, PROT_READ | PROT_WRITE, MAP_SHARED, fdSrc, 0);

Is write access really necessary?

how to interpret output file

Hi,

I've successfully run the VVP and got the output result (Hooray!), just wondered how should I interpret the output file, and what does each column mean in the output?

Thanks,

gnomAD vvp non-existant + vcf gives segmentation error

Since there is no gnomAD vvp build anymore I tried to run the example build (which works in the example provided), it gives segmentation error in my own vcf file which I formatted as it is in the example. What should I do to fix this problem.

how to generate percentile scores

I have successfully run vvp on my vcf file, but the output does not contain percentile scores. How do I compile percentile scores for my gene of interest using CRD curves, as done in the publication? I can't find it in the documentation or in the additional files of the publication.

Thanks,
Odette

request: bioconda

Hi
Great tool - any chance you can or will get this up as a package in bioconda?
Would be great for the bioinformatics and NGS community

Best

Steve

how to use background based on gnomAD ?

hi
I I've successfully run the VVP,but I do know how to use background based on gnomAD? I download gnomad.062717.build.tar.gz and unzip it ，can not find gnomad.062717.build.bin and gnomad.062717.build.max.
can you provide some tips or command to build gnomAD background?

WARNING: chromosome not in offsets

Hi Steven,
I tried running vvp on a test sample data and but it gives 'WARNING: chromosome chrY not in offsets' error all chr, also I have not been able to figure what the CSQ parameter is about. If you clarify these, it would be helpful for my analysis.

command : ./VVP -i /test_sample.vcf -d /VVP-pub-master/gnomad.062717.build -v CSQ,4,6,1,15 1>test_sample.out
error : WARNING: chromosome chrY not in offsets

can't build with gcc

I was able to build this with clang, but building with gcc (v 5.4.0) fails with the messages below.

CMakeFiles/VVP.dir/score_variants.c.o: In function `main':
score_variants.c:(.text+0xf1c): undefined reference to `gzdopen'
score_variants.c:(.text+0xf39): undefined reference to `gzopen'
score_variants.c:(.text+0x1076): undefined reference to `gzgets'
score_variants.c:(.text+0x111a): undefined reference to `gzgets'
CMakeFiles/VVP.dir/score_variant.c.o: In function `compute_score':
score_variant.c:(.text+0x175): undefined reference to `log'
score_variant.c:(.text+0x198): undefined reference to `log'
score_variant.c:(.text+0x1ab): undefined reference to `log'
score_variant.c:(.text+0x1ce): undefined reference to `log'
score_variant.c:(.text+0x1e1): undefined reference to `log'
CMakeFiles/VVP.dir/score_variant.c.o:score_variant.c:(.text+0x204): more undefined references to `log' follow
CMakeFiles/VVP.dir/parse_vcf.c.o: In function `parse_allele_frequency_line':
parse_vcf.c:(.text+0x4661): undefined reference to `gsl_rng_env_setup'
parse_vcf.c:(.text+0x4668): undefined reference to `gsl_rng_taus'
parse_vcf.c:(.text+0x4670): undefined reference to `gsl_rng_alloc'
parse_vcf.c:(.text+0x46a7): undefined reference to `gsl_rng_uniform_int'
parse_vcf.c:(.text+0x47ad): undefined reference to `gsl_rng_uniform_int'
parse_vcf.c:(.text+0x48ad): undefined reference to `gsl_rng_uniform_int'
parse_vcf.c:(.text+0x49a4): undefined reference to `gsl_rng_uniform_int'
parse_vcf.c:(.text+0x4a7a): undefined reference to `gsl_rng_free'
collect2: error: ld returned 1 exit status
CMakeFiles/VVP.dir/build.make:250: recipe for target 'VVP' failed
make[2]: *** [VVP] Error 1
CMakeFiles/Makefile2:67: recipe for target 'CMakeFiles/VVP.dir/all' failed
make[1]: *** [CMakeFiles/VVP.dir/all] Error 2
Makefile:83: recipe for target 'all' failed
make: *** [all] Error 2