Jim's installation notes

Revision as of 23:10, 13 December 2007 by JimHu (Talk | contribs)

Jump to: navigation, search

Back to Chado New Users

Installation on PPC mac

  • First do this on my G4 laptop. Then consider whether to install on the XServe.

Installation section of Getting Started describes how to get Chado from CVS. Note, when updating, be sure to be at the directory level just above schema. Getting Started refers to the INSTALL document




Comes installed on OS X


BioPerl is currently not installed on my laptop, as I ran into the dreaded gd installation problems. Let's try again. Installing BioPerl on Unix applies to Mac OSX once the prereqs are done (I think). Some incomplete notes here from installing on the server. As I recall, this failed the last time I tried on my laptop. However, I think Matías Giovannini updated his Installation guide for gd since the last time I tried it. With the new instructions gd went smoothly. notes here.

To get BioPerl from cvs see here

GO perl

  • The url provided in the Chado install document is not working for me. Contacted Mike Cherry, who provided an alternative url: Downloaded (this URL will change with their next update, use the other to browse the available downloads from the GO consortium).
  • go_perl is only one module inside the utilities. From its install docs, there are more CPAN modules to install:
    • install GO::Parser
      got this warning:
      NOTICE: ** You do not have XML::Parser::PerlSAX installed **

      This module is not required for go-perl, but it is required if you wish to parse Obo-XML files. It is also required by the go-db-perl library for loading the GO Database.

      If you do not intend to use either Obo-XML or go-db-perl, then you may continue. Even if you do care about these, you may continue and install XML::Parser::PerlSAX at some future time.
    • I'm going to need the XML::Parser at some point, I suspect. I recall having problems with this during BioPerl installation. install XML::Parser::PerlSAX follows dependency to install XML::Parser, which complains about not having expat.
      • installing expat: get it from http://sourceforge.net/projects/expat/
      • install XML::Parser. Note that quitting and restarting CPAN helps if an installation already failed once.
      • install XML::Parser::PerlSAX
    • install GraphViz - also optional, but likely to be useful

Curiously, the install instructions suggest that this is all I need to do for go_perl. So despite downloading the tarball, all I needed was the CPAN installation??

Other Perl Dependencies

install Bundle::GMOD

As is often the case with the CPAN modules, watching the Terminal window shows lots of tests failing. Also, it looks for modules I'm pretty sure I've already installed.

Failed during this command:
 GAAS/libwww-perl-5.805.tar.gz                : make_test NO
 MIYAGAWA/Class-DBI-Pager-0.08.tar.gz         : make_test NO
 CAPTTOFU/DBD-mysql-4.003.tar.gz              : make_test NO
 TJMATHER/XML-DOM-1.44.tar.gz                 : make_test NO
 JROBINSON/SQL-Translator-0.08.tar.gz         : make_test NO
 TMTM/Class-DBI-v3.0.16.tar.gz                : make_test NO
 ADAMK/AppConfig-1.64.tar.gz                  : make_test NO
 ABW/Template-Toolkit-2.18.tar.gz             : install NO
 DMAKI/Class-DBI-Pg-0.08.tar.gz               : make_test NO
 SZABGAB/Spreadsheet-ParseExcel-0.28.tar.gz   : make_test NO
 BYRNE/SOAP-Lite-0.69.tar.gz                  : make_test NO
 MIROD/XML-DOM-XPath-0.13.tar.gz              : make_test NO

Set up default gmod location

mkdir /usr/local/gmod

Is that it? I don't feel like this is right. Since I installed bundle GMOD using CPAN, I'm not sure where it's gone. Probably /Library/Perl... Forge ahead anyway.

Install chado

A few more prereq issues

 perl Makefile.PL 
The following enviroment variables not detected:
Please read the INSTALL document and set required environment variables.

Oops. Went back and edited .profile

export GMOD_ROOT="/usr/local/gmod" 
export CHADO_DB_NAME="chado_trial" 

also did these on command line, since .profile settings don't refresh until opening a new Terminal (I think).

 perl Makefile.PL

Lots of output. Failed. Possible culprits:

The envirnment variable GO_ROOT is not set.
Ontology loading may not work correctly without it.

Extracting /Users/jimhu/schema/chado/bin/../lib/Chado/AutoDBI.pm (with variable substitutions)
Can't locate LWP/Simple.pm in @INC (@INC contains: lib /System/Library/Perl/5.8.6/darwin-thread-multi-2level /System/Library/Perl/5.8.6 /Library/Perl/5.8.6/darwin-thread-multi-2level /  
Library/Perl/5.8.6 /Library/Perl /Network/Library/Perl/5.8.6/darwin-thread-multi-2level /Network/Library/Perl/5.8.6 /Network/Library/Perl /System/Library/Perl/Extras/5.8.6/darwin-
thread- multi-2level /System/Library/Perl/Extras/5.8.6 /Library/Perl/5.8.1 .) at lib/Chado/Builder.pm line 15.
BEGIN failed--compilation aborted at lib/Chado/Builder.pm line 15. 
Compilation failed in require at load/Build.PL line 3.
BEGIN failed--compilation aborted at load/Build.PL line 3.
sh: line 1: ./Build: No such file or directory
unable to create Build from load/Build.PL:

Added the line to .profile to set GMOD_ROOT (this was in the install doc and I must have missed it earlier).

export GO_ROOT=$HOME/go-dev

I assume this was set up during install of GO::Parser??

cpan>install LWP::Simple 

failed tests. I think these are based on failing to connect to localhost via apache...it may reflect something about OSX that should be set in the config. Try forcing it.

Makefile.pl etc.

Do the Makefile.pl again. End of output is

Warning: the following files are missing in your kit:
Please inform the author.
Warning: prerequisite DBIx::DBStag 0.07 not found.
Checking if your kit is complete...
Looks good
Writing Makefile for Bio-Chaos
Writing Makefile for gmod

Makefile written.  Now you should do the following, in order:
  1. make              (creates necessary build files)
  2. sudo make install (creates $GMOD_ROOT and subdirectories)
  3. make load_schema (loads SQL schema into database)
  4. make prepdb      (loads basic data)
  5. make ontologies  (loads data for various ontologies)

Optional Targets:
  make html         (auto-generates html documentation for each module)
  make images       (generates images of the schema in several formats)
  make rm_locks     (removes ontology lock files, allowing installation
                     of ontologies on successive builds of the database
                     without removing the ontology files altogether)
  make clean        (remove build related files and ontology tmp dir)
  make instructions (at any moment display these instructions)


Followed steps 1-5. Chose to load all the ontologies.

make ontologies

./Build ontologies

Available ontologies:
[1] Relationship Ontology
[2] Sequence Ontology
[3] Gene Ontology
[4] Chado Feature Properties
[5] Cell Ontology
[6] Plant Ontology

Which ontologies would you like to load (Comma delimited)? [0]  1,2,3,4,5,6
fetching files for Relationship Ontology
   loading...sh: line 1: stag-storenode.pl: command not found
System call 'stag-storenode.pl -d 'dbi:Pg:dbname=chado_trial;host=localhost;port=5432' --user jimhu  --password lambda  tmp/obo/OBO_REL/rel.oboxml' failed: 32512
Died at lib/Chado/Builder.pm line 315, <STDIN> line 1.
make: *** [ontologies] Error 2

Back to cpan.

install DBIx::DBStag

failed. Force it. The output during force is more informative. Need to create a role for root in postgres...or is it supposed to work with my username, which has a role? After force, make ontologies seems to work without creating a role for root.

Getting lots of warnings while loading SO. example:

HINT:  Use the escape string syntax for backslashes, e.g., E'\\'.
WARNING:  nonstandard use of \\ in a string literal
LINE 1: ... (cvterm_id, value, type_id, rank) VALUES ('222', 'gene\\:<g...

GO is taking a long time...not surprising. It would be nice if there was some kind of progress indicator. Went to group meeting upon return, seems like all loaded, but:

DBD::Pg::db do failed: ERROR:  duplicate key violates unique constraint "cvterm_c1"
DBD::Pg::db do failed: ERROR:  duplicate key violates unique constraint "cvterm_c1"

This has been observed before and is noted in the install docs.

make html


Can't locate SQL/Translator.pm

Back to CPAN. install SQL::Translator failed. Force it. During make test, saw (excerpts)

       all skipped: Missing dependency: Text::RecordParser
       all skipped: Missing dependency: Spreadsheet::ParseExcel

Could have sworn an Excel parser was already installed. I don't think I need that, but Text::RecordParser sounds useful. CPAN installed it successfully (along with yet another dependency).

After that, make html works. As for using it...

make images


Can't load producer 'GraphViz' : Error loading GraphViz as SQL::Translator::Producer::GraphViz

Wait a minute! I know I installed a GraphViz earlier today.

cpan[12]> install SQL::Translator::Producer::GraphViz
SQL::Translator::Producer::GraphViz is up to date (1.12).

It's a different GraphViz. I still need GraphViz.

cpan>install GraphViz

identifies a C library dependency, graphviz. Shades of gd. Follow http://www.graphviz.org/ and find that there are Mac OS X package installers, which is a good thing, since graphviz has many other dependencies of its own. Download 2.12 revision 1 for PPC and install. graphviz installs into /usr/local as /usr/local/graphviz-2.12. Make symlink.

ln -s /usr/local/graphviz-2.12 /usr/local/graphviz
export PATH=/usr/local/graphviz/bin:$PATH

Try installing the CPAN Graphviz again. Need to export in that session, but then it installs, following dependencies. Do make images again. Takes a long time. The good news is that it runs. The bad news is that it:

  • creates image files in the wrong place - they're showing up in the directory where chado was downloaded.
  • the image files are nonfunctional. They're showing up as 0K.

I wonder if there's a versions problem with graphviz.

make rm_locks and make clean

I think they worked.

make instructions


make: *** No rule to make target `instructions'.  Stop.

No idea how to deal with this one, but it shouldn't stop me from going on.

Load the yeast sample data

Per instructions: get the yeast file from ftp://ftp.yeastgenome.org/pub/yeast/data_download/chromosomal_feature/saccharomyces_cerevisiae.gff. cd to my downloads directory and do

gmod_bulk_load_gff3.pl --organism yeast --gfffile saccharomyces_cerevisiae.gff

gmod_bulk_load_gff3.pl was installed in /usr/bin, so it can be run from anywhere. Messages during load (with comments/questions):

(Re)creating the uniquename cache in the database... 
Creating table...

What table is being created? Chado installation already created a zillion tables.

Populating table...
Creating indexes...Done.
Preparing data for inserting into the chado_trial database
(This may take a while ...)

This GFF file has CDS and/or UTR features that do not belong to a 
'central dogma' gene (ie, gene/transcript/CDS).  The features of 
this type are being stored in the database as is.

Should be interesting to look at these.

Loading data into feature table ...
Loading data into featureloc table ...
Loading data into feature_relationship table ...
Loading data into featureprop table ...
Loading data into feature_cvterm table ...
Loading data into synonym table ...
Loading data into feature_synonym table ...
Loading data into dbxref table ...
Loading data into feature_dbxref table ...
Skipping analysisfeature table since the load file is empty...

No BLAST reports etc. Loading sequences (if any) ...

Optimizing database (this may take a while) ...
  (feature featureloc feature_relationship featureprop feature_cvterm synonym feature_synonym dbxref feature_dbxref analysisfeature ) Done.

While this script has made an effort to optimize the database, you
should probably also run VACUUM FULL ANALYZE on the database as well

Set up Gbrowse to work with Chado

Set up www user

psql chado_trial

Altered the instructions, since I'm pretty sure that OSX uses username www instead of nobody for apache. Actually I'm not sure the username matters as long as it matches the name in the gbrowse conf file, at least for gbrowse.

 GRANT SELECT ON feature_synonym      TO www;
 GRANT SELECT ON synonym              TO www;
 GRANT SELECT ON feature_dbxref       TO www;
 GRANT SELECT ON dbxref               TO www;
 GRANT SELECT ON feature              TO www;
 GRANT SELECT ON featureloc           TO www;
 GRANT SELECT ON cvterm               TO www;
 GRANT SELECT ON feature_relationship TO www;
 GRANT SELECT ON cv                   TO www;
 GRANT SELECT ON gffatts              TO www;
 GRANT SELECT ON feature_cvterm       TO www;
 GRANT SELECT ON feature_gcontext     TO www;
 GRANT SELECT ON gcontext             TO www;
 GRANT SELECT ON featureprop          TO www;
 GRANT SELECT ON pub                  TO www;
 GRANT SELECT ON feature_pub          TO www;
 GRANT SELECT ON db                   TO www;

The two GRANT statements for feature_gcontext and gcontext gave errors, since these tables don't exist. Additional GRANTs may be needed if Gbrowse is adapted to display additional views.

Make a gbrowse conf file

Copy the example file from gbrowse.

sudo cp /Library/WebServer/Documents/gbrowse/contrib/conf_files/07.chado.conf chado_yeast.conf

Edit it.

Gbrowse displays everything except the gbrowse image. Check the error logs.

DBD::Pg::st execute failed: ERROR:  permission denied for relation analysisfeature

fix with:

GRANT SELECT ON analysisfeature                  TO www;
Features from yeast sample GFF imported to Chado

It works!!!