🚀 go-pugleaf

All,

Here is the March newsletter, ok, ok, so it's April but at least it's early
April.

Ed
-- 
 ------------------------------------------------------------------------
| Ed Griffiths, Acedb development, Informatics Group,                    |
|        Wellcome Trust Sanger Institute, Wellcome Trust Genome Campus,  |
|               Hinxton, Cambridge CB10 1SA, UK                          |
|                                                                        |
| email: edgrif@sanger.ac.uk  Tel: +44-1223-494780  Fax: +44-1223-494919 |
 ------------------------------------------------------------------------


----------------------------------------------------------------------------
----------------------------------------------------------------------------

ACEDB User Group Newsletter - March 2003
########################################

If you want to have this newsletter mailed to you _or_ you want to make
comments/suggestions about the format/content then send an email to
acedb@sanger.ac.uk.

-----------------------------------------------

This month there are improvements/additions to giface commands, cut/pastable
popup error messages (an AceDB first...), long awaited FMap keyboard
shortcuts, control of timestamping frequency for server log files, a gene
feature finder executable and fix for possibly the most irritating FMap bug
in the world.

-----------------------------------------------

General News
************

We seem to be back on track with monthly releases and newsletters. The SMap
code has been used extensively by the C. Elegans group at Sanger for mapping
homologies, alleles and other features and seems stable and working well.
We'll be producing a write up of SMap tags and other associated tag sets
(e.g. the homology tag set) that are supported by AceDB for next month or
sooner if possible.

-----------------------------------------------

New Features
************

New and improved giface commands
++++++++++++++++++++++++++++++++

If you use giface then you may be interested to note the following
changes/additions to the "smap" and "seqget" commands available from the
"gif" submenu in giface.

"smap": The smap command has been made much faster so that you should be
able to use it to find the position of a child smap'd feature in any parent
of that feature quickly enough to service web server requests if you are
using the AceDB server version of giface.

For example you can find the position of genes on forward and reverse strand
of a chromosome in the worm:

     acedb> gif smap -from sequence:m7.1
     SMAP Sequence:CHROMOSOME_IV 11082185 11081556 1 444 INTERNAL_GAPS
     // 0 Active Objects
     acedb>

     acedb> gif smap -from sequence:m7.2
     SMAP Sequence:CHROMOSOME_IV 11084106 11085837 1 1551 INTERNAL_GAPS
     // 0 Active Objects
     acedb>

Here "m7.1" is on the reverse strand of CHROMOSOME_IV from bases 11081556 to
11082185, while "m7.2" is on the forward strand from bases 11084106 to
11085837. "INTERNAL_GAPS" indicates that these features are sets of exons
and hence have gaps

If you specify a coordinate range and the -from object is outside of that
coordinate range the smap call will return as its last word "OUTSIDE_AREA"
to indicate that the object could be mapped but was outside the area of
interest.

"seqget": The seqget command now has a "-noclip" option. With this option,
when you use seqget specifying a coordinate range, you can decide whether
features should be clipped to the coordinate range or, if you specify
"-noclip" whether features that overlap should be mapped to their full
extent.

For example:

   =================|=========================|=================== sequence
                   100                       200
                _________            ______
               /         \          /      \
           ====           ==========        ========= feature

                    <-------- clip ----------->

           <------------------- noclip ------------->

Here we have asked to see features from bases 100 to 200 of the sequence,
the default (clipped) will return only the part of the feature that does not
extend past this range, the noclip option returns the full features.

Clipping is useful for something like fMap where a user is only interested
in what features look like in a particular range, the unclipped option is
useful when you want to see the full features that overlap a particular
range perhaps to then output them as in GFF or some other format.

FMap and display of reverse strand features.
++++++++++++++++++++++++++++++++++++++++++++

If you are in tree display of a sequence object and click on a child
sequence that is on the reverse strand feature the fMap will now
automatically be displayed showing the reverse strand. Previously the object
would be on the left/up-strand of the fMap display and so may not even have
been displayed if its method did not include the "Show_up_strand" tag.

You saw it here first
+++++++++++++++++++++

OK, once again AceDB is ahead of the game....

How many times have you seen some error in a program that has been popped up
in a helpful little dialog box and what you'd really like to do is cut/paste
that message into an email and send if off with some choice vituperative
comments of your own to the programs developers. Of course what you discover
is that you can't cut/paste the message so you have to copy it out verbatim
and then you make a mistake and accidentally leave out the vital piece of
information and, and, and....

Well in AceDB you can cut/paste the message, just move your mouse pointer
over the message, you will see it go highlighted, click on it with the left
button and then use the middle button to cut/paste it into your email.

AceDB, ahead of the game...well, OK, sometimes we are...

FMap keyboard shortcuts
+++++++++++++++++++++++

After an infinity of time and an almost as large number of people
"volunteering" to put some shortcuts into the fMap something has finally
happened. Here is a list of newly added keyboard shortcuts in the FMap:

left,right arrows        creep left/right
Ctrl left,right arrows   page left/right

up,down arrows           creep up/down
Ctrl up,down arrows      page up/down

=,-                      zoom little in/out
Ctrl =,-                 zoom lots in/out

Shift up,down arrows     cycle through highlighting features
                         column by column

Page Up/Down             page up or down
Home/End                 go to start/end of current view

spacebar                 recalculate
Enter/Return             pick (like mouse click on current item)
"r"                      reverse-complement
"w"                      whole (zoom fMap to whole sequence)

Controlling the frequency of timestamps in the server log file
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

A while ago I reported an AceDB server performance bug caused by time
stamping every single log file message, at the time a quick fix was made to
allow the database administrator to select time stamping on an "all or
nothing" basis. You can now choose intermediate levels of time stamping
using the new TIMESTAMP_FREQUENCY keyword in the wspec/serverconfig.wrm file
of the database. You can use this keyword to say how often you would like
time stamps to appear in the server logfile.

>From the wspec/serverconfig.wrm description:

//////////////////////////////////////////////////////////////////
// TIMESTAMP_FREQUENCY:
// If the TIMESTAMP_FREQUENCY flag is uncommented in this file
// then the number following it will be used to set the frequency
// with which log messages are timestamped.
//
// The timestamp frequency will be interpreted as:
//
// 0   no timestamping.
// 1   timestamp every message.
// n   timestamp every n messages.
//
// The default is 0.
//
//TIMESTAMP_FREQUENCY 0

GFF output, unknown fields
++++++++++++++++++++++++++

When AceDB outputs features from an FMap in GFF format, there may be fields
that it doesn't have data for, e.g. the "source" field. AceDB used to output
the string "unknown" for these fields, but this is not distinctive enough as
curators sometimes use this string as the value for GFF_source tag in their
methods. In future AceDB will output the string "*UNKNOWN*" for any field in
a GFF record for which it cannot establish a value, hence you will be able
to distinguish between features that you want to call "unknown" and actual
missing data in the database.

-----------------------------------------------

Articles
********

New Gene Features program available
+++++++++++++++++++++++++++++++++++

(This article is courtesy of Rob Clack rnc@sanger.ac.uk & Kevin Howe
klh@sanger.ac.uk)

You can now download a separate executable which uses the Phil Green/LaDeana
Hillier's gene finder program to search for localised gene finder features.

The standalone genefeatures program scans a FASTA-format sequence for
localised gene features which it then writes to stdout in GFF format.

To achieve this it uses a series of example profiles, both positive examples
of the feature of interest and matching "pseudo" examples (eg pieces of
sequence which have some splice-site characteristics, but are not
splice-sites.)

The program was originally based on Phil Green and LaDeana Hillier's
GENEFINDER program, some bits of which have been incorporated into fMap.

Thanks to Kevin Howe for help/programming.

-----------------------------------------------

Bugs Fixed
**********

Losing the active feature on scrolling in FMap
++++++++++++++++++++++++++++++++++++++++++++++

An old and frustrating bug for annotators has been fixed. It used to be that
if you selected a feature and then in scrolling up/down that feature went
off the screen, FMap would forget that you had selected the feature, so then
you had to click it again. Now you will find that FMap can just about always
remember which feature you selected, the only exception being if you scroll
so far up/down that the feature is no longer in the calculated section of
sequence that FMap is displaying.

Server exit message twice
+++++++++++++++++++++++++

A minor bug: the server used to output its exit message to the server log
twice, this is now fixed.

-----------------------------------------------

Developers Corner
*****************

Incredible dotter bug
+++++++++++++++++++++

Fixed an incredible bug in dotter:

Here's how xqseqDisp is declared:

static char xqseqDisp[3][MAXALIGNLEN+1] ;

Here's how xqseqDisp is initialised:

    for (i = 0; i < MAXALIGNLEN; i++)
        qseqDisp[i] =   sseqDisp[i] =
            xqseqDisp[0][i] = xqseqDisp[2][i] = xqseqDisp[3][i] =
                qseqDispCrick[i] = sseqDispCrick[i] = 0;

Youch, we go clean off the end of the xqseqDisp array with xqseqDisp[3][i],
thanks to the Alpha compiler for spotting that one buried in the code.

Array and keyset bug
++++++++++++++++++++

Thanks to Rob for sorting this one out:

Fixed bug where key is inserted at posn 0 in array, but code thought index
of -1 meant array was empty. Changed to use arrayMax to verify array size.

New graph package call
++++++++++++++++++++++

A small additional function has been added to the graph package, you can now
set the size of graphs of any type specifying pixels for the size using
graphRawBounds().

Some changes to the gex/graph package
+++++++++++++++++++++++++++++++++++++

To implement keyboard shortcuts the keyboard callback routines had to be
changed to allow the keyboard modifier (Ctrl, shift etc) to be passed back
as well as the key. The GTK layer now passes back modifier keys that are
pressed at the same time as normal keys. This allows the graph package to
pass such key presses through to FMap or other graph package based displays.

Note that where possible these actions are dealt with at the GTK level, in
FMap for instance, horizontal scrolling can be dealt with at the GTK level,
as horizontal scrolling is via a GTK widget. Vertical scrolling is dealt
with at the graph/FMap level because it involves the FMap type scroll bar.
By dealing with these simpler user interface actions at the GTK level we
make them generally available to all displays that use GTK type widgets for
scrolling etc.

Some changes to the build system
++++++++++++++++++++++++++++++++

The AceDB home directory is now a little less full as all build log files
are now held in a .BUILDLOGS subdirectory. If a build fails look in here for
the log files to see what went wrong.

Some changes to smap to fix a pernicious bug
++++++++++++++++++++++++++++++++++++++++++++

By default smap clips the coordinates of mapped features to the coordinate
range specified, this resulted in a bug in sMapMapEx(). sMapMapEx() needs to
know the true length of the parent, not the clipped length. The length is
needed so that features that are reversed with respect to the parent and
hence must be positioned from the end of the parent can be positioned
correctly. To do this the SMapKeyInfoStruct has to hold the length of the
parent.

The function now also checks to see if a feature is outside the area of the
smap and return the new SMAP_STATUS_OUTSIDE_AREA code for this. We need this
because its possible for a feature to be mapped correctly but still be
outside the specified coordinate range for a particular mapping.

-----------------------------------------------

March monthly build now available.
**********************************

You can pick up the monthly builds from:

Sanger users
     ~acedb/RELEASE.DEVELOPMENT
External users
     http://www.acedb.org/Software/Downloads/monthly.shtml

-----------------------------------------------

Next User Group Meeting - D319, 3.00pm, Thursday, 10th April 2003
*****************************************************************

----------------------------------------------------------------------------
----------------------------------------------------------------------------
Ed Griffiths <edgrif@sanger.ac.uk>
Last modified: Fri Apr 4 13:51:43 BST 2003