Hi all.
I want to visualize hapmap data (on site hapmap project) using Gbrowse. I have download all data that Gbrowse need on hapmap site.
I also performed all step following Readme file (bellow), of course, i changed follow me (ie: name database, description ...). and all the step did'nt get any error. but when i access with firefox browse, it no get anything.
What should i do to slove it.
Readme's content
#---------------------------------
# HAPMAP GBROWSE
#---------------------------------
This guide and the contents of this directory are intended to help users install their
local copy of the HapMap GBrowse feature viewer.
### Note on hardware requirements
Loading the GFF-files into Bio:
B::GFF requires a substantial amount of memory (1.5-2Gb). Also,
running the MySQL server with the resulting database giving acceptable performance requires a fairly
powerful machine. Make sure you have sufficient hardware before attempting to install the HapMap
GBrowse locally. A typical workstation will almost certainly not be enough, a server-class machine
with at least 4Gb RAM is a bare minimum.
### Get data
Download all available GFFs from the HapMap website:
ncftpget ftp://www.hapmap.org/gbrowse/latest/gff/*gff.gz
This will get you the following data from the HapMap project:
gt+allele_freqs_*.gff.gz : genotyped SNP features, with allele and genotype frequencies
recomb_hotspots.gff.gz : recombination hotspot features
recomb_rate.gff.gz : recombination rate features
and bincounts for SNP and gene features, in low (500Kb) and high (5-20Kb) resolution windows:
*density*.gff.gz
and various annotations from external sources. For all feature types, see GBrowse help pages for
more information:
Download FASTA-files for the genome assembly from UCSC (NB command gets B35 assembly used by the
current HapMap release (rel#21), adjust to get more recent assemblies). Note that even if you do
not plan to use the DNA-sequence in your local installation, the HapMap frequency glyph require it.
ncftpget ftp://hgdownload.cse.ucsc.edu/apache...es/chr*.fa.zip
### Get software
Install the latest GBrowse/Bio:B::GFF package, and make sure you have a working installation
by installing some of the test data bundled with GBrowse (as described in tutorial).
Download and install custom HapMap glyphs & plugins into your GBrowse, and make sure
they do not have compilation errors due to missing libraries in your local machine etc.
Install the latest Bioperl distribution, so you definately have working GFF3 handling
in the Bio:B::GFF library (there was a buggy version in circulation a while ago).
Download a patched Bio:B::GFF bulkloader script from the HapMap website and add to
your existing Bioperl installation. The patch does the following:
-forces the attributes of genotyped SNP features for each population to 'pile up'
on a single fdata table entry, instead of the default behaviour of creating seperate
features per SNP per population.
-enables decompression of zipped (.zip) FASTA files from UCSC on the fly.
### Load feature data into MySQL
To get around 'full-table' problems with the fattribute_to_feature table, make sure you have at least MySQL 5.0.6
so you can have tables >4Gb without modifying server parameters or table properties (see
http://dev.mysql.com/doc/refman/5.0/en/full-table.html).
To load GFFs, a two-step procedure is needed to get around a Perl memory problem in the bulk loader. The problem
may not exist on newer Unix platforms & Perl distributions, so you can try just loading the entire set of GFFs
in one fell swoop with the bulk loader:
-Load part I: run the bulkloader on all GFFs *except* the dbSNP file, plus the dir containing the FASTA-files.
This takes 6-7 hours on our machine:
nohup time bulk_load_gff -c -d dbi:mysql:[dbname] --user [user name] --password [XXXX]
--gff3_munge --maxfeature 10000000000 --local --fasta dna/* *.gff* &
-Load part II: run the regular loader on the dbSNP files to add them to the bulkloaded data. This takes
8-10 hours on our machine:
nohup time load_gff --dsn dbi:mysql:[dbname] -gff3_munge --local dbsnp_b124_on_B34.gff3.gz &
[NOTE: You can combine both the steps if you are reasonably confident of the dataload procedure]
Set proper permissions on the database, so the user running Apache (typically 'nobody') can read it:
mysql>grant select on [dbname].* to nobody;
mysql>flush privileges;
Download the HapMap GBrowse config, copy to your config dir, modify and test:
Finally, if everything is working, optimize MySQL database tables to get better performance. This
can take an hour or even more:
mysql>optimize table, fattribute, fattribute_to_feature, fdata,fdna,fgroup,fmeta,ftype;
-------------------------------
[email protected]
So it
I want to visualize hapmap data (on site hapmap project) using Gbrowse. I have download all data that Gbrowse need on hapmap site.
I also performed all step following Readme file (bellow), of course, i changed follow me (ie: name database, description ...). and all the step did'nt get any error. but when i access with firefox browse, it no get anything.
What should i do to slove it.
Readme's content
#---------------------------------
# HAPMAP GBROWSE
#---------------------------------
This guide and the contents of this directory are intended to help users install their
local copy of the HapMap GBrowse feature viewer.
### Note on hardware requirements
Loading the GFF-files into Bio:
B::GFF requires a substantial amount of memory (1.5-2Gb). Also,running the MySQL server with the resulting database giving acceptable performance requires a fairly
powerful machine. Make sure you have sufficient hardware before attempting to install the HapMap
GBrowse locally. A typical workstation will almost certainly not be enough, a server-class machine
with at least 4Gb RAM is a bare minimum.
### Get data
Download all available GFFs from the HapMap website:
ncftpget ftp://www.hapmap.org/gbrowse/latest/gff/*gff.gz
This will get you the following data from the HapMap project:
gt+allele_freqs_*.gff.gz : genotyped SNP features, with allele and genotype frequencies
recomb_hotspots.gff.gz : recombination hotspot features
recomb_rate.gff.gz : recombination rate features
and bincounts for SNP and gene features, in low (500Kb) and high (5-20Kb) resolution windows:
*density*.gff.gz
and various annotations from external sources. For all feature types, see GBrowse help pages for
more information:
Download FASTA-files for the genome assembly from UCSC (NB command gets B35 assembly used by the
current HapMap release (rel#21), adjust to get more recent assemblies). Note that even if you do
not plan to use the DNA-sequence in your local installation, the HapMap frequency glyph require it.
ncftpget ftp://hgdownload.cse.ucsc.edu/apache...es/chr*.fa.zip
### Get software
Install the latest GBrowse/Bio:
B::GFF package, and make sure you have a working installationby installing some of the test data bundled with GBrowse (as described in tutorial).
Download and install custom HapMap glyphs & plugins into your GBrowse, and make sure
they do not have compilation errors due to missing libraries in your local machine etc.
Install the latest Bioperl distribution, so you definately have working GFF3 handling
in the Bio:
B::GFF library (there was a buggy version in circulation a while ago).
Download a patched Bio:
B::GFF bulkloader script from the HapMap website and add toyour existing Bioperl installation. The patch does the following:
-forces the attributes of genotyped SNP features for each population to 'pile up'
on a single fdata table entry, instead of the default behaviour of creating seperate
features per SNP per population.
-enables decompression of zipped (.zip) FASTA files from UCSC on the fly.
### Load feature data into MySQL
To get around 'full-table' problems with the fattribute_to_feature table, make sure you have at least MySQL 5.0.6
so you can have tables >4Gb without modifying server parameters or table properties (see
http://dev.mysql.com/doc/refman/5.0/en/full-table.html).
To load GFFs, a two-step procedure is needed to get around a Perl memory problem in the bulk loader. The problem
may not exist on newer Unix platforms & Perl distributions, so you can try just loading the entire set of GFFs
in one fell swoop with the bulk loader:
-Load part I: run the bulkloader on all GFFs *except* the dbSNP file, plus the dir containing the FASTA-files.
This takes 6-7 hours on our machine:
nohup time bulk_load_gff -c -d dbi:mysql:[dbname] --user [user name] --password [XXXX]
--gff3_munge --maxfeature 10000000000 --local --fasta dna/* *.gff* &
-Load part II: run the regular loader on the dbSNP files to add them to the bulkloaded data. This takes
8-10 hours on our machine:
nohup time load_gff --dsn dbi:mysql:[dbname] -gff3_munge --local dbsnp_b124_on_B34.gff3.gz &
[NOTE: You can combine both the steps if you are reasonably confident of the dataload procedure]
Set proper permissions on the database, so the user running Apache (typically 'nobody') can read it:
mysql>grant select on [dbname].* to nobody;
mysql>flush privileges;
Download the HapMap GBrowse config, copy to your config dir, modify and test:
Finally, if everything is working, optimize MySQL database tables to get better performance. This
can take an hour or even more:
mysql>optimize table, fattribute, fattribute_to_feature, fdata,fdna,fgroup,fmeta,ftype;
-------------------------------
[email protected]
So it