Interval data structure
Switch branches/tags
Nothing to show
Clone or download
Latest commit ec83c5a Apr 16, 2018

README.md

Build Status

GIGGLE is a genomics search engine that identifies and ranks the significance of shared genomic loci between query features and thousands of genome interval files.

For questions and discussion about GIGGLE please visit/join the mailing list: https://groups.google.com/d/forum/giggle-discuss

For more information about how GIGGLE works, please read the manuscript in Nature Methods: https://www.nature.com/articles/nmeth.4556

Or watch a presentation about GIGGLE on YouTube (14m 37s) GIGGLE

Usage

GIGGLE has two high-level functions:

  • index creates an index from a directory of bgzipped annotations (BED files or VCF files)
  • search takes a region or a file of regions and searches them against an index
giggle, v0.6.3
usage:   giggle <command> [options]
     index     Create an index
     search    Search an index

Indexing

giggle, v0.6.3
usage:   giggle index -i <input files> -o <output dir> -f
         options:
             -s  Files are sorted
             -i  Files to index (e.g. data/*.gz)
             -o  Index output directory
             -f  For reindex if output directory exists

Searching

giggle, v0.6.3
usage:   giggle search -i <index directory> [options]
     options:
         -i giggle index directory
         -r <regions (CSV)>
         -q <query file>
         -o give results per record in the query file (omits empty results)
         -c give counts by indexed file
         -s give significance by indexed file (requires query file)
         -v give full record results
         -f print results for files that match a pattern (regex CSV)
         -g genome size for significance testing (default 3095677412)
         -l list the files in the index  

Example

To demonstrate GIGGLE indexing and searching, we will curate and query a genome repeat reference dataset.

This example will use gargs from https://github.com/brentp/gargs

wget -O gargs https://github.com/brentp/gargs/releases/download/v0.3.8/gargs_linux
chmod +x gargs

or if you are on a Mac

wget -O gargs https://github.com/brentp/gargs/releases/download/v0.3.8/gargs_darwin
chmod +x gargs

This reference will be based on the following annotation from the UCSC genome browser:

  • Repeat Masker
  • Segmental Duplications
  • Microsatellites
  • Simple Repeats

UCSC stores data as tables, and the relevant columns vary between files, so we must take some care in curating the data.

mkdir repeat
url="http://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/rmsk.txt.gz"
curl -s $url | gunzip -c | cut -f 6,7,8,11,12,13 > repeat/rmsk.bed

url="http://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/simpleRepeat.txt.gz"
curl -s $url | gunzip -c | cut -f 2,3,4,17 > repeat/simpleRepeat.bed

url="http://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/microsat.txt.gz"
curl -s $url | gunzip -c | cut -f 2,3,4 > repeat/microsat.bed

url="http://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/genomicSuperDups.txt.gz"
curl -s $url | gunzip -c | cut -f 2,3,4,5 > repeat/genomicSuperDups.bed

url="http://hgdownload.soe.ucsc.edu/goldenPath/hg19/database/chainSelf.txt.gz"
curl -s $url | gunzip -c  | cut -f 3,5,6,7,10,11 > repeat/chainSelf.bed

Once all of the data are in bed files in the repeat directory, we sort and bgzip the files then index.

mkdir repeat_sort
giggle/scripts/sort_bed "repeat/*.bed" repeat_sort 4
giggle index -i "repeat_sort/*gz" -o repeat_sort_b -f -s

A GIGGLE index can be queried and the output formatted in a variety of ways. The most basic is to search a single interval and get the number of overlaps for each database file.

giggle search -i repeat_sort_b -r 1:200457776-200457776

#repeat_sort/chainSelf.bed.gz   size:1058543    overlaps:0
#repeat_sort/genomicSuperDups.bed.gz    size:51599  overlaps:0
#repeat_sort/microsat.bed.gz    size:41572  overlaps:0
#repeat_sort/rmsk.bed.gz    size:5298130    overlaps:1
#repeat_sort/simpleRepeat.bed.gz    size:962714 overlaps:0

To search only a subset of database files use the -f option, which takes a comma separated list of regular expressions. Only those database files that match one of the regular expressions will be considered.

giggle search -i repeat_sort_b -r 1:200457776-200457776 -f rmsk,simple

#repeat_sort/rmsk.bed.gz    size:5298130    overlaps:1
#repeat_sort/simpleRepeat.bed.gz    size:962714 overlaps:0

To retrieve the original records for each overlap, use the -v option. This is useful for detailed filtering and summaries.

giggle search -i repeat_sort_b -r 1:200457776-200457776 -f rmsk,simple -v

chr1    200457488   200457811   L2a LINE    L2  repeat_sort/rmsk.bed.gz

GIGGLE also accepts query files in either bed.gz or vcf.gz formats. When a query file is given all of the above options are valid. In addition, GIGGLE can perform statistical tests between the query file and each database file using the -s option. These tests include the:

  • odds ratio that estimates the enrichment of observed v. expected
  • the Fisher's two tailed, left tailed, and right tailed tests that estimate p-values
  • the GIGGLE combo score that combines the odds ratio and Fisher's two tailed tests
giggle search -i repeat_sort_b -q bed.bed.gz -s
#file                             file_size   overlaps              odds_ratio         fishers_two_tail       fishers_left_tail   fishers_rigth_tail       combo_score
repeat_sort/chainSelf.bed.gz        1058543     410556  1.4753028915472957e-10  9.9350208733579337e-201 9.9350208733579337e-201   1                        0
repeat_sort/genomicSuperDups.bed.gz   51599      31434  1.2707669134541957      7.63990775297919e-91    1                         3.8346413983550421e-91  31.153365312726858
repeat_sort/microsat.bed.gz           41572      12320  1.0772570268110226      3.7685521408880773e-10  0.99999999982079036       1.9339512340673677e-10   1.0117655469639908
repeat_sort/rmsk.bed.gz             5298130    1599051  1.0068476528982573e-10  3.9822462776218267e-200 2.2545997326354446e-200   1                        0
repeat_sort/simpleRepeat.bed.gz      962714     308460  1.0977857352008243e-10  4.4719559780220419e-201 4.4719559780220419e-201   1                        0

Original records can also be retrieved and grouped by query interval with the -v -o options.

giggle search -i repeat_sort_b -q ALL.wgs.integrated_sv_map_v2.20130502.svs.genotypes.vcf.gz -v -o | head

####1     645710  ALU_umary_ALU_2 A       <INS:ME:ALU>    .       .       TSD=null;SVTYPE=ALU;MEINFO=AluYa4_5,1,223,-;SVLEN=222;CS=ALU_umary;AC=35;AF=0.00698882;NS=2504;AN=5008;EAS_AF=0.0069;EUR_AF=0.0189;AFR_AF=0;AMR_AF=0.0072;SAS_AF=0.0041;SITEPOST=0.9998
chr1    18392   804926  chr19   60000   244029  repeat_sort/chainSelf.bed.gz
chr1    70007   667633  chr6    60000   147635  repeat_sort/chainSelf.bed.gz
chr1    114546  672699  chr1    260861  676115  repeat_sort/chainSelf.bed.gz
chr1    126221  713929  chr7    56425213        56942114        repeat_sort/chainSelf.bed.gz
chr1    130987  703609  chr7    55804351        56479927        repeat_sort/chainSelf.bed.gz
chr1    131111  751608  chr1    222641449       224228809       repeat_sort/chainSelf.bed.gz
chr1    227417  706973  chr5    14937   199599  repeat_sort/chainSelf.bed.gz
chr1    228316  648877  chr6    60000   193267  repeat_sort/chainSelf.bed.gz
chr1    230805  804926  chr16   60000   164433  repeat_sort/chainSelf.bed.gz

Building

Dependencies

From a fresh install of Ubuntu, the following steps should provide all the required dependencies.

sudo apt install gcc make autoconf zlib1g-dev libbz2-dev libcurl4-openssl-dev libssl-dev ruby

Giggle command line interface

git clone https://github.com/ryanlayer/giggle.git
cd giggle
make
export GIGGLE_ROOT=`pwd`
cd ..

Run tests

The first set of tests require bedtools to be in your path.

sudo apt install g++ python

git clone https://github.com/arq5x/bedtools2.git
cd bedtools2
make
cd bin
export PATH=$PATH:`pwd`
cd ../..

Now run the tests

cd $GIGGLE_ROOT/test/func
./giggle_tests.sh
cd ../unit
make
cd ../../..

Hosted data and services

Data

Roadmap Epigenomics: https://s3.amazonaws.com/layerlab/giggle/roadmap/roadmap_sort.tar.gz

UCSC Genome browser: https://s3.amazonaws.com/layerlab/giggle/ucsc/ucscweb_sort.tar.gz

Fantom5: https://s3.amazonaws.com/layerlab/giggle/fantom/fantom_sort.tar.gz

Interactive heatmap

http://ryanlayer.github.io/giggle/index.html?primary_index=ec2-54-227-176-15.compute-1.amazonaws.com/rme&ucsc_index=ec2-54-227-176-15.compute-1.amazonaws.com/ucsc

Web server (optional)

This is based on libmicrohttpd

mkdir -p $HOME/usr/local/
wget http://ftpmirror.gnu.org/libmicrohttpd/libmicrohttpd-0.9.46.tar.gz
tar zxvf libmicrohttpd-0.9.46.tar.gz
cd libmicrohttpd-0.9.46
./configure --prefix=$HOME/usr/local/
make
make install

export LD_LIBRARY_PATH=$HOME/usr/local/lib/

cd ..

sudo apt install libtool

wget https://github.com/json-c/json-c/archive/json-c-0.12.1-20160607.tar.gz
tar xvf json-c-0.12.1-20160607.tar.gz  
cd json-c-json-c-0.12.1-20160607
./configure --prefix=$HOME/usr/local/
make
make install

cd $GIGGLE_ROOT
make
make server
cd ..

To host the site shown in Supplemental Figure 3, you will need host web servers for both the Roadmap Epigenomics data and the UCSC data. Here we will run both servers on the same host from ports 8080 and 8081 and access the web services using localhost, but these are general steps and apply to many other configurations including hosting the data sets on different servers.

wget https://s3.amazonaws.com/layerlab/giggle/roadmap/roadmap_sort.tar.gz
tar -zxvf roadmap_sort.tar.gz

# NOTE, if the following command gives "Too many open files" try:
ulimit -Sn 16384
$GIGGLE_ROOT/bin/giggle index -s -f \
    -i "roadmap_sort/*gz" \
    -o roadmap_sort_b 
    
wget https://s3.amazonaws.com/layerlab/giggle/ucsc/ucscweb_sort.tar.gz
tar -zxvf ucscweb_sort.tar.gz

$GIGGLE_ROOT/bin/giggle index -s -f \
    -i "ucscweb_sort/*gz" \
    -o ucscweb_sort_b

Start a web server for each index.

$GIGGLE_ROOT/bin/server_enrichment -i roadmap_sort_b/ -u /tmp/ -d $GIGGLE_ROOT/examples/rme/data_def.json -p 8080 &
$GIGGLE_ROOT/bin/server_enrichment -i ucscweb_sort_b/ -u /tmp/ -d $GIGGLE_ROOT/examples/ucsc/data_def.json -p 8081 &

If you get Access-Control-Allow-Origin errors, then pass the `-a` option to `server_enrichment`

Pass these two services to the web interface through URL arguments:

http://ryanlayer.github.io/giggle/index.html?primary_index=localhost:8080&ucsc_index=localhost:8081

These data are also being served here:

http://ryanlayer.github.io/giggle/index.html?primary_index=ec2-54-227-176-15.compute-1.amazonaws.com/rme&ucsc_index=ec2-54-227-176-15.compute-1.amazonaws.com/ucsc

Example analysis

NOTE: Index files and query files MUST be bgzipped (https://github.com/samtools/htslib, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3042176/).

Roadmap Epigenomics

# details of how to recreate the data at 
# https://github.com/ryanlayer/giggle/blob/master/examples/rme/README.md
wget https://s3.amazonaws.com/layerlab/giggle/roadmap/roadmap_sort.tar.gz
tar -zxvf roadmap_sort.tar.gz

# NOTE, if the following command gives "Too many open files" try:
# ulimit -Sn 16384
$GIGGLE_ROOT/bin/giggle index -s -f \
    -i "roadmap_sort/*gz" \
    -o roadmap_sort_b 

wget ftp://ftp.ncbi.nlm.nih.gov/geo/samples/GSM1218nnn/GSM1218850/suppl/GSM1218850_MB135DMMD.peak.txt.gz
# take the just the top peaks
zcat GSM1218850_MB135DMMD.peak.txt.gz \
| awk '$8>100' \
| cut -f1,2,3 \
| $GIGGLE_ROOT/lib/htslib/bgzip -c \
> GSM1218850_MB135DMMD.peak.q100.bed.gz

# List files in the index
$GIGGLE_ROOT/bin/giggle search -l \
    -i roadmap_sort_b/ 

# Search
$GIGGLE_ROOT/bin/giggle search -s \
    -i roadmap_sort_b/ \
    -q GSM1218850_MB135DMMD.peak.q100.bed.gz \
> GSM1218850_MB135DMMD.peak.q100.bed.gz.result


# Plot
sudo apt install python python-pip python-tk
pip install matplotlib
$GIGGLE_ROOT/scripts/giggle_heat_map.py \
    -s $GIGGLE_ROOT/examples/rme/states.txt \
    -c $GIGGLE_ROOT/examples/rme/EDACC_NAME.txt \
    -i GSM1218850_MB135DMMD.peak.q100.bed.gz.result \
    -o GSM1218850_MB135DMMD.peak.q100.bed.gz.result.3x11.pdf \
    -n $GIGGLE_ROOT/examples/rme/new_groups.txt \
    --x_size 3 \
    --y_size 11 \
    --stat combo \
    --ablines 15,26,31,43,52,60,72,82,87,89,93,101,103,116,120,122,127 \
    --state_names $GIGGLE_ROOT/examples/rme/short_states.txt \
    --group_names $GIGGLE_ROOT/examples/rme/new_groups_names.txt

APIs

Python by Brent Pedersen

from giggle import Giggle
index = Giggle('existing-index-dir') # or Giggle.create('new-index-dir', 'files/*.bed')
print(index.files)

result = index.query('chr1', 9999, 20000)
print(result.n_files)
print(result.n_total_hits) # integer number sum of hits across all files

print(result.n_hits(0)) # integer number of hits for the 0th file...

for hit in result[0]:
    print(hit) # hit is a string

Installation

make sure you have liz, libcurl, libcrypto, libbz2 and liblzma installed in the appropriate place on your system.

git clone --recursive https://github.com/brentp/python-giggle
cd python-giggle
python setup.py test
python setup.py install

Go by Brent Pedersen

GoDoc

import (
    giggle "github.com/brentp/go-giggle"
    "fmt"
) 

func main() {

    index := giggle.Open("/path/to/index")
    res := index.Query("1", 565657, 567999)

    // all files in the index
    index.Files()

    // int showing total count
    res.TotalHits()

    // []uint32 giving number of hits for each file
    res.Hits()

    var lines []string
    # access results by index of file.
    lines = res.Of(0)
    fmt.Println(strings.Join(lines, "\n"))
    lines = res.Of(1)
}

Docker

giggle-docker by Ryuichi Kubo

docker run kubor/giggle-docker giggle -h
Unknown command
giggle, v0.6.3
usage:   giggle <command> [options]
         index     Create an index
                  search    Search an index