FAQ

The EMBOSS Frequently Asked Questions list
==========================================

Last modified: 8th July 2009


SECTION A - Help and Support
============================

Q) How do I report bugs?

A) mail emboss-bug@emboss.open-bio.org


Q) How do I contact the core development team?

A) Rather than email their personal addresses, for EMBOSS matters we
   request that you use the address:
   
   emboss-bug@emboss.open-bio.org

   This will ensure that your email will be seen by the appropriate
   developer and that it will get logged with our tracking system.


Q) Are there any mailing lists about EMBOSS?

A) emboss@emboss.open-bio.org 
   is an open list (anyone can join) for general announcements and
   discussions by users. 

   emboss-dev@emboss.open-bio.org 
   is a closed list (subscription requests have to be approved) for
   discussions by developers using EMBOSS. 
   
   emboss-announce@emboss.open-bio.org 
   is an open list for announcements of new releases. 
   
You can access the archives, subscribe/unsubscribe and alter
the way email is sent to you (e.g. digests) by visiting:

  http://emboss.open-bio.org/mailman/listinfo/emboss
  http://emboss.open-bio.org/mailman/listinfo/emboss-dev
  http://emboss.open-bio.org/mailman/listinfo/emboss-announce


Q) Where's the documentation?
   
A) All the documentation can be found at the following web page
   http://emboss.sourceforge.net/


Q) Where's the application documentation?

A) http://emboss.sourceforge.net/apps/

   and in the EMBOSS distribution, installed under
   share/EMBOSS/doc/programs/html (HTML files) and
   share/EMBOSS/doc/programs/text (plain text, as used by the tfm program).

   You can also use the EMBOSS application called tfm to display the plain
   text documentation of another application e.g.:

   % tfm seqret


Q) Is there a tutorial?

A) See the EMBOSS tutorial on the website:
   http://emboss.sourceforge.net/docs/emboss_tutorial/


Q) Is there a quick guide?

A) There is a dated quick guide provided with the source code as
   the file doc/manuals/emboss_qg.pdf. A postscript file is
   also available. However, the primary source of up-to-date
   information is the web site. The quick guide was not written by
   the EMBOSS team and it is therefore not maintained by them.


Q) Is there a table of substitutes for GCG programs?

A) Several lists are available, but not maintained by the EMBOSS
   team. For example:
   http://www.biobind.com/faq/emboss/gcg-emboss.html


Q) Citation: Is there a reference I can cite for EMBOSS?

A) Rice,P. Longden,I. and Bleasby,A.
   "EMBOSS: The European Molecular Biology Open Software Suite"
   Trends in Genetics June 2000, vol 16, No 6. pp.276-277




SECTION B - Installation & Compilation
======================================

Q) How do I compile EMBOSS?

A) If this is the first time trying to compile all you need to do is:

   ./configure
   make
 
   The above will produce the EMBOSS programs in the 'emboss'
   subdirectory and you can set your PATH variable to point there.
   This method is suitable for EMBOSS developers.

   For system-wide installations we recommend installing the
   EMBOSS programs into a different directory from the source
   code (e.g. in the directory tree /usr/local/emboss). To do
   this type (e.g.):

      ./configure --prefix=/usr/local/emboss
      make                    [to make sure there are no errors, then]
      make install            [if there are no errors]

   You should then add (e.g.) /usr/local/emboss/bin to your PATH
   variable:
      set path=(/usr/local/emboss/bin $path)        [csh/tcsh shells]
      export PATH="$PATH:/usr/local/emboss/bin"     [sh/bash shells]


If you wish to re-compile the code at any stage then you should type
"make clean" and then configure and make source again as appropriate.


Q) I have a Linux system and compilation ends prematurely saying that
   it can't find the -lX11 libraries ... but I know I have X11
   installed.

A) This should not happen with versions of EMBOSS later than 6.0.1
   as the configuration will inform you of missing files and
   terminate at that stage.
     
   If you have a version of EMBOSS <= 6.0.1 then the problem
   is that you probably have the X11 server installed but you haven't
   installed the X11 development files. For example, on RPM distributions,
   you need to install:
     xorg-x11-proto-devel           (xorg-based X11 distros) or
     XFree86-devel                  (XFree86-based X11 distros)

   After installing those system files you will need to:
     make clean
   and re-perform the './configure' command from the top EMBOSS
   source directory.


Q) I'm trying to compile EMBOSS with PNG support

A) Your system will need to have zlib (www.zlib.net: current version
   is 1.2.3), libpng (www.libpng.org: current version is 1.2.35) and
   gd (www.libgd.org: current version is 2.0.35).

   The version of GD must be >= 2.0.28 if it is to be used with EMBOSS.

   Note that the above packages often come with your operating system
   distribution but may not be installed by default i.e. they
   are optional packages which you can install at a later date.  

   Also note that, as well as having the above system libraries installed,
   you must also have installed their associated development files.
   (these typically have 'devel' in the package names on Linux systems).
   So, on Linux RPM systems you would need to make sure that
   packages called names similar to "gd" and "gd-devel" are installed.

   MacOSX users can often find the above packages available on the
   MacPorts site (www.macports.org).

   However, for some operating systems, there may be no freeware
   sites where you can find pre-compiled versions of zlib/libpng/gd
   and you may have to compile one or more of them from their
   source distributions (URLs given above).

   You can unpack the tar.gz or tar.bz2 files in any directory, and
   install them in a common area.

   By default everything (including EMBOSS) installs
   in /usr/local.


   To install, pick up the sources and then:

   gunzip -c zlib-1.2.3.tar.gz   | tar xf -
   gunzip -c libpng-1.2.35.tar.gz | tar xf -
   gunzip -c gd-2.0.35.tar.gz     | tar xf -

   cd zlib-1.2.3
   ./configure
   make
   make install
   ./configure --shared
   make
   make install
   cd ..


   cd libpng-1.2.35
   ./configure
   make
   make install
   cd ..


   cd gd-2.0.35
   ./configure --without-freetype --without-fontconfig
   make
   make install
   cd ..

   To compile with the local version your EMBOSS ./configure line should now
   read:-
   ./configure --with-pngdriver=/usr/local


Q) When installing EMBOSS recently I get a load of errors due to
   libraries not found.  The main problem is that I have an old version of
   libz but no libgd in my system libraries and EMBOSS looks there first
   to try to locate these libraries.  I have the correct versions installed
   elsewhere. Are there any suggestions for setting the library search path
   or am I missing something really obvious?

A) There are the
	   --without-pngdriver    
   and
           --with-pngdriver=dir
   flags. Did you try them? If the libraries are in /opt/png/lib then
   set "dir" to /opt/png i.e. one level above the "lib" directory.


Q) Can I get the latest code via CVS ?

A) Yes. Here is the information you will need:-

   Make sure you have cvs on your system. Then log into the cvs server at
   'open-bio.org' as: user 'cvs' with password 'cvs'. 

   cvs -d :pserver:cvs@cvs.open-bio.org:/home/repository/emboss login
   The password is 'cvs'

   To checkout the EMBOSS source code tree, put yourself in a local directory
   just above where you want to see the EMBOSS directory created. For example
   if you wanted EMBOSS to be seen as /home/joe/src/emboss... then cd into
   /home/joe/src then checkout the repository by typing: 

   cvs -d :pserver:cvs@cvs.open-bio.org:/home/repository/emboss checkout emboss

   Or if you want to update a previously checked-out source code tree:
   cvs -d :pserver:cvs@cvs.open-bio.org:/home/repository/emboss update

   You can logout from the CVS server with:
   cvs -d :pserver:cvs@cvs.open-bio.org:/home/repository/emboss logout

   (this is a read only server). 


Q) How do I compile the CVS (developers) version?

A) You will need automake, autoconf, (GNU) make and libtool for this.

   The gcc compiler is recommended. The host cc compiler should work.
   (Note that some non-gcc compilers may need '--include-deps' added to
   the automake command line.)


   Then type:

   autoreconf -fi
   ./configure            # --enable-warnings is a useful developers' switch
   make

   For more info on the configurability of the build try 
   ./configure --help
  

Q) Can you give an example of how to install an EMBASSY package?

A) Here is how to install PHYLIP given the various ways you can install
   the main EMBOSS package.

   a) from the anonymous cvs code.
    
       1) Go to the phylip directory
       cd embassy/phylipnew

       2) make the configuration file
       autoreconf -fi

       3) configure and compile
       ./configure (use same options as you used to compile emboss,
                    especially the --prefix))
       make
       make install

    b) from PHYLIP-3.6b.tar.gz 

    available from our FTP server ftp://emboss.open-bio.org/pub/EMBOSS/
    in file PHYLIP-3.6b.tar.gz 


    If you have done a full installation of EMBOSS using a 'prefix'
    e.g. you configured with ./configure --prefix=/usr/local/emboss
    and followed this with a 'make install' (highly recommended) then:

       1) gunzip and untar the file anywhere
       gunzip PHYLIP-3.6b.tar.gz
       tar xf PHYLIP-3.6b.tar

       2) go into the phylip directory
       cd PHYLIP-3.6b

       3) configure and compile
       ./configure (use same options as you used to compile emboss,
                    especially the --prefix))
       make
       make install

       N.B. If you configured without using a prefix but did do
            a 'make install' (or specified a prefix of /usr/local,
            which amounts to the same thing) then you must configure
            using:
            ./configure --prefix=/usr/local --enable-localforce

    If, on the other hand, you did not do a 'make install' of EMBOSS
    then:

       1) Go to the emboss directory 
       cd EMBOSS-6.0.1

       2) make new directory embassy if it does not exist already.
       mkdir embassy

       3) Go into that directory
       cd embassy

       4) gunzip and untar the file
       gunzip PHYLIP-3.6b.tar.gz
       tar xvf PHYLIP-3.6b.tar

       5) go into the phylip directory
       cd PHYLIP-3.6b

       6) configure and compile
       ./configure (use same options as you used to compile emboss)
       make

       7) Set your PATH to include the full path of the 'src'
          directory


Q) Can I alter the location of the ACD files?

A)  EMBOSS will find the acd files in the install directory or in the
    original source directory (depending on how you configured it) so
    there is usually no reason to change this. However,
    should you wish to do so then it can either be done
    in the emboss.default (or ~/.embossrc) file:

      env emboss_acdroot /usr/local/share/EMBOSS/acd

    or by setting an environment variable on the command line:

      setenv EMBOSS_ACDROOT /usr/local/share/EMBOSS/acd


Q) What benefits do I gain from using the associated (EMBASSY) versions of
    software.

A)  a) You can read any sequence type that EMBOSS can handle.

    b) the associated software will use the EMBOSS acd files so the
    naming of output/input files is taken care of and will check all
    values before the program is run. Command line arguments are used
    instead interactive menu based ones.


Q) I have EMBOSS installed on our development server and I'm preparing a
    dispatch which will send it out to about 20 remote sites. 

    I ran the configure with the --prefix option to install to a private
    directory.  I also collected all the data files (rebase, transfac etc.) to
    another directory and extracted the information from them with the relevant
    programs.

    My intention was to simply transfer the EMBOSS install directory and the
    Data directory to all sites, using symlinks where necessary so that the
    directory paths corresponded.

    However, when testing this I have found a couple of problems;

    1) Although the EMBOSS programs work, I can't see any of the extracted
    data.
    For instance remap gives the error;

               EMBOSS An error in remap.c at line 167:
            Cannot locate enzyme file. Run REBASEEXTRACT

    This is despite the fact that I have both the EMBOSS install and the Data
    directories in the same place as on the development machine (which works).

    2) The other major problem is that I can no longer see my databases defined
    in emboss.default.  Again, the file exists, and is in the same place as on
    the development machine, but the box it is transferred to gives an empty
    list from showdb.

    Does anyone know where EMBOSS stores the information about the location of
    these files?  It can't have installed anything outside the original
    installation directory (wasn't installed as root), so I'm guessing that the
    problem stems from the program resolving symlinks at some point.

A) It is inside the binaries ...

    EMBOSS 'knows' the location of the files because it is picked up during the
    configure, when you build your copy, and included in the binaries.

    You can see it during compilation, especially of ajnam.c (where it is used):

    -DAJAX_FIXED_ROOT=\"/full/source/path\" -DPREFIX=\"/install/prefix/path\" 

    To copy binaries, you need to define environment variable(s) to
    override the compile-time definitions, unless you can make the path
    (e.g.  /usr/local) the same for the installations at each site. 

    emboss.default can set environment variables too, but you need to tell
    EMBOSS where to find that file. 
         setenv EMBOSS_ROOT /dir/for/default/file

    and then, in the emboss.default file you can set:
         SET EMBOSS_ACDROOT /install/dir/share/EMBOSS/acd

    or (this overrides it) you can use another environment variable:
         setenv EMBOSS_ACDROOT /install/dir/share/EMBOSS/acd


Q) I have downloaded the EMBOSS source and installed it for use at XYZ
   University without any difficulty.  The administration guide has advice
   on configuring the software using emboss.default, and has some examples for
   allowing access to SRS indexes.  That appears to be done via the program
   getz, which is not part of the EMBOSS package. 

A) If you have SRS installed (so you have local SRS index files) you will
   have a local copy of the getz program, which is part of SRS.

   If you do not have SRS, you can build your own index files using dbxflat,
   dbxgcg (if you have GCG), dbiblast (if you have blast) and dbxfasta. This is
   the usual solution for sites that have no other database indexing in use.

   You can also use SRS servers remotely, to get single entries, using their
   URLs. No extra software is needed (EMBOSS just uses the HTTP protocol).

   Of course, if you really need to build your own SRS indexes you
   could install it. SRS is a commercial product, but academic licences are
   available.


Q) I am not getting full static files even when I configure with
   --disable-shared

A) This most often happens when using GNU LD. If both shared and
   static versions of a library exist then GNU LD will take the
   shared library as preferred. The root of this problem is libtool.
   You can, however, force complete static images by adding a
   definition to your "make" line: 

   make "LDFLAGS=-Wl,-static"


SECTION C - Graphics
====================


Q) What Graphics options are available?

A) To see what graphics drivers are available type ? at the prompt for
   the graph type and this will give you a list.

   Here are some of those:-

   ps      -> Postscript
   cps     -> Colour Postscript
   x11     -> X display. (also called xterm and xwindows)
   hpgl    -> HP Laserjet III, HPGL emulation mode.
   png     -> PNG (you will need libpng, zlib and gd for this)
   gif     -> GIF (you will need libpng, zlib and gd for this)
   tek     -> Tektronix Terminal
   none    -> None.
   data    -> Writes out points to a file for graphs.
   meta    -> plplot meta file.


Q) Plotters and pen colours.
    
A) The hp drivers presume the pens are loaded as:-

    SP1 Red
    SP2 Yellow
    SP3 Green
    SP4 Blue
    SP5 Magenta
    SP6 White/Off
    SP7 Black
    SP8 Cyan

    Otherwise your output will have different colours.


Q) Browsing the Gnu site I came across libplot, libxmi and plot utils. 
   Would these be a suitable replacement for PLPLOT?

A) So far they are still GPL licensed rather than LGPL.

   Robert Maier promised a while back to LGPL them but has not yet. A pity 
   ad it makes all the difference for linking in third party applications to
   EMBOSS. 


Q) I get error messages when I try to display X11 on my PC.
   I am running the Hummingbird Exceed X11 emulator.

A) This should only affect versions of EMBOSS pre-5.0.0
   The Hummingbird Exceed X11 emulator (and maybe other systems) use the
   'TrueColor' display by default. 
   You should change the configuration settings so that it uses 'PseudoColor'.
   In version 6.1 of Hummingbird Exceed, this is done by clicking on the
   Exceed section of the Windows 98 tool bar at the bottom of the screen.
   Select 'Tools', then 'Configuration', then Screen Definition.
   A window will appear.
   In the 'Server Visual' popup menu on the left, select 'PseudoColor'.
   Click on 'OK'.


SECTION D - Data Files
======================


Q) How do I use my own private data file?

A) You may wish to amend one of the standard EMBOSS data files.  Some of
  the data files you might wish to alter are the translation table files. 

'transeq' has been written to only read in one of the standard
translation files:

  EGC.0
  EGC.1
  EGC.2
  EGC.3
  EGC.4
  EGC.5
  EGC.6
  EGC.9
  EGC.10
  EGC.12
  EGC.11
  EGC.13
  EGC.14
  EGC.15

These files are the only ones that you can specify to 'transeq'. 

If you wish to create your own specialised translation table, then you
should pick one of them to amend. 

For instance you may decide that you will use the file EGC.15 as you
would never want to otherwise use it. 

Use the program 'embossdata' to get a copy of this file:

% embossdata -fetch -filename EGC.15
Finds or fetches the data files read in by the EMBOSS programs
File '/fu/bar/emboss/data/EGC.15' has been copied successfully.

Edit the file EGC.15 to suit your requirements.

Specify '-table 15' when you run 'transeq' to use this altered file. 
'transeq' will then look for the file 'EGC.15' and will find it in your
current directory before it finds the default one in the EMBOSS_DATA
directory. It will therefore use your local copy.

You may get confused with many copies of changed files floating about. 
To check which copy of a file is being used (the default EMBOSS_DATA
one or a potential local copy) use 'embossdata':

% embossdata -filename EGC.15
Finds or fetches the data files read in by the EMBOSS programs
# The following directories can contain EMBOSS data files.
# They are searched in the following order until the file is found.
# If the directory does not exist, then this is noted below.
# '.' is the UNIX name for your current working directory.

File ./EGC.15                                            Exists
File .embossdata/EGC.15                                  Does not exist
File /home/joe/EGC.15                                    Does not exist
File /home/joe/.embossdata/EGC.15                        Does not exist
File /usr/local/emboss/data/EGC.15                       Exists

This shows that a copy of EGC.15 exists in your current directory and so
will be used in preference to the default one in the EMBOSS_DATA
directory. 


SECTION E - Sequence Files and Formats
======================================

Q) How do you write sequences to different files instead of writing them
   all to one file?

A) EMBOSS is not good at writing multiple sequences to different files.
   You could try using 'nthseq' to pull out one sequence at a time.

   You should consider using the '-ossingle' qualifier.  This writes
   sequences to separate FASTA files, but the file names depend on the
   sequence ID.  This is rarely used.  It is there because -ossingle is the
   default for GCG output format.  The output filename is currently ignored
   when ossingle is used, and the filename depends on each individual
   sequence.


Q) What sequence formats are supported?

A) Tens of them(e.g.):

   gcg, embl, swissprot, fasta, ncbi, genbank, nbrf, codata, strider,
   clustal, phylip, acedb, msf, ig, staden, text, raw, asis


Q) What is the difference between TEXT and RAW formats?

A) TEXT accepts everything in the sequence file as sequence.
   RAW accepts only alphanumeric and whitespace characters and rejects
   anything else. 


Q) What is ASIS format?

A) The "filename" is really the sequence.
   This is a quick and easy way of reading in a short fragment of
   sequence without having to enter it into a file. 

   For example:
   % program -seq asis::ATGGTGAGGAGAGTTGTGATGAGA




Q) I have some very short protein sequences that EMBOSS thinks are
   nucleic sequences.  How do I force EMBOSS to treat them as nucleic acid
   sequences? 

   For example:
   > cat seq1
   A
   > cat seq2
   I

   % water seq1 seq2 -stdout -auto
   Smith-Waterman local alignment.
   An error has been found: Sequence is not nucleic

   Here, 'water' automatically (and wrongly) thinks that A is adenosine
   instead of alanine and fails when it reads in seq2 and expects to
   find another nucleic acid sequence - but 'I' is not a valid base and
   so it fails. 

A) For many sequence formats there is no way to specify the sequence
   type in the file, so EMBOSS has to guess.

   There is a flag that can force EMBOSS programs to treat sequences as
   nucleic or protein. 

   'water -help -verbose'
   shows the full list of sequence qualifiers.

   If you follow the sequence USA with '-sprotein' EMBOSS will check
   that it is a valid protein sequence and then treat it as such. 

   If you need to force a sequence to be DNA, the qualifier is
   '-snucleotide'

   The qualifier must follow the sequence to apply to one sequence, or
   can go at the start of the command line to refer to all sequences, for
   example:

   'water -sprotein seq4 seq3 -stdout -auto'

   You can also use '-sprotein1' anywhere on the command line to refer
   to the first sequence and '-sprotein2' to refer to the second sequence. 

   Of course, like all EMBOSS qualifiers, you can shorten them so long
   as they are still unique.  In this case, '-sp' and '-sn' will work
   (or '-sp1' and '-sp2' if you need the numbers). 


SECTION F - Limits and contraints
=================================

Q) Is there a maximum size for sequences?

A) The maximum size for any program depends only on how much memory your
   system has.

   Of course, some programs (and some program options) can take up too
   much memory, or simply run very slowly. 

   You might have a constraint imposed on your usage of memory.
   Try using the Unix command 'limit' to look at such constraints.
   Try using the Unix command 'unlimit' or 'ulimit' to remove the constraints,
   eg:
     % unlimit stacksize
     % unlimit vmemoryuse


Q) GCG has a somewhat arbitrary fragment length limit of 2500 bp for gel*.
   Is there a similar limit for MSE under EMBOSS?

A) No. MSE has no limit, you are only limited by how much memory you
   have. 


SECTION G - Running under Web servers and from scripts
======================================================


Q) I am trying to write a web interface for an emboss program
   and run apache. The program complains that there is no HOME
   directory set.

A) Just put these at the top after your 'use CGI' ( whatever) statements.

   $ENV{HOME}=<web owner root>  i.e. /usr/local/apache
   $ENV{EMBOSS_DATA}=<emboss data directory>

   These two are important, but you can also pass other "constants". 


SECTION H - Programs and options
================================

Q) Plotting with pepwheel gives interesting output.
   pepwheel -turns=8 -send=30 sw:p77837 -auto
   gives a helical wheel plot but the residues are plotted so every two
   circles are sat on top of one another. 

A) That is the correct answer.

   Instead of 3.6 residues per turn, (5 turns in 18 steps) you seem to
   have a helix with 8 turns in 18 steps (4 in 9).

   Try -turns=4 -steps=9 ... but only if you are sure that is the 
   way your helix goes.

   We could play with checking for common factors in pepwheel,
   but we don't know of any biologically meaningful setting that would
   cause problems.


Q) In prettyplot, how do you specify an output file name for the plot file?

A) prettyplot -auto ~/wordtest/globin-nogap.msf -graph ps
   Creates prettyplot.ps
   The name is generated automatically.  To set this to something more
   discriptive use -goutfile .i.e

   prettyplot -auto ~/wordtest/globin-nogap.msf -graph ps -goutfile=hello
   Creates hello.ps


Q) I'm using the editor 'mse', but I don't know how to save my edited sequences
   at the end of the editing session. 

A) Use Control-Z to get into command mode.
   Then use the SAVE command which will prompt for the file name.


Q) I would like to know if EMBOSS can perform nucleotide contig assembly
   similar to the function that GCG gelproject/gelview has.  And if yes, is
   there any size restriction on the number of basepair and the number of
   contigs? thank you. 

A) EMBOSS does not cover the sort of contig assembly you describe. We
   have provided an EMBOSS wrapper for the MIRA package.