Image: The fingerprint image analysis system

Image is a package of analysis algorithms for processing gel images from restriction digest fingerprinting experiments.

Image has been tightly integrated with a friendly user interface and provides a robust tool for large scale physical mapping. Image is able to process gels from a wide variety of scanning technologies and has been tested on various fingerprinting protocols, producing normalized band and gel images as output.

The software package has evolved over many years since it first saw the light of day on a VAX machine in the shape of analysis programs written by John Sulston et.al.

[Genome Research Limited]

Overview

What does Image do?

Image extracts a normalized band pattern for each lane (clone) on a gel. To achieve this several procedures are run on each gel in turn. The cycle starts when the user scans the film image or otherwise imports gel image data. The cycle ends when the normalized band position are transferred to a database and FPC takes over and matches up the clones to assemble contigs.

The Image analysis is comprised of these parts:

  • Lane tracking - Lanes are recognizedautomatically. A grid is superimposed on the gel image. The lanes may be edited manually by easy drag & drop functions.
  • Band calling - An analysis module traces the band pattern along the lanes and tries to identify bands. This stage still involves frequent human interaction. An easy-to-use editor helps the user to scan quickly through the gel and clear up ambiguities. Bands may be added, deleted and moved by mouse or keyboard.
  • Marker locking - In order to compare band patterns from one gel to another, all band positions are normalized to one master gel. A set of DNA fragments of known length is loaded as a marker lane every nth lane with sample lanes in between. This standard pattern is matched to the pattern from the master gel. The editor is used to get the best match between the two sets of bands (match between the standard lane and each marker lane on the current gel).
  • Normalisation - Once the marker lane patterns are locked onto the standard lane, the band positions of the sample lanes in between are normalized - so that each lane appears to have been run on the master gel with all distortions due to differences in temperature, etc. cancelled out.

A suite of helper programs is provided to further automate the analysis process. Scripts take care of transfer and copying operations - the user does not have to know the details of the file system once installation is complete.

How to obtain Image

We supply pre-compiled binaries for supported platforms.

Getting help

If you encounter any problems when installing or running Image, online help is available. You can also contact the support team for help.

Summary

After down loading and installing and installing the software package, these are all the separate programs that are available to perform various tasks. Some of them are started from the command line, others are called automatically in the background.

The tutorial will explain step by step how the programs are started and operated in order to analyze an example gel.

imscan
script to scan autoradiograms on an Amersham film reader® using the scanning software installed in ~scanner/
imauto
script to run the analysis cycle in batch mode (executes automatic analysis modules one by one on a list of gels)
imimport
script used to create an Image gel project from an existing image file, it creates all directories and copies the image file across
im3conv
program to convert almost any image file format (including Amersham .dat files, TIFF images and ABI® gelfiles) to Image's own private format - the hidump format; this is done automatically when the gel is first opened up
im3
core of the system - this is the interactive project management, control and gel editing system, it calls the automatic modules and displays the results graphically and provides manual edit functions for each stage of the analysis
pathfinder
automatic module : creates the lane map and writes the lane data for amStep3
amStep3
automatic module : creates traces by scanning along each lane on the raw image and then finds bands along each lane, it writes the traces and the band data for step amStep4
amStep4
automatic module : takes each marker lane and matches it to a standardized pattern of the standard lane. The number of bands in each marker lane have to equal the number of lanes in the standard lane. Bands are matched up in order and the normalization factors used in amStep5 are written.
amStep5
automatic module : called when a gel is finished - normalizes all raw band positions using the distortion factors calculated by amStep4, it writes the normalized bands and also writes a normalized trace image, scanned along each sample lane
imtransfer
script to copy the normalized bands and the normalized trace image to its place in the fpc database, where the gel data can be imported into an fpc project

Conventions

Types of input

The Image processing software current deals with three types of input:

Gel format

In order for Image to correctly differentiate between marker and sample lanes your gel MUST have the following format

  • a marker lane is loaded every n-th lane
  • the first and last lane are marker lanes
  • there is always an equal number of samples loaded between markers
  • the marker repeat parameter is the number of samples between two markers plus ONE
  • there are usually 6 samples loaded between markers, hence the marker repeat parameter is 7
  • if you use autoradiograms as input, it is advisable to draw a line on the loading well to aid the well finding algorithm in the band calling module
  • all further annotation on the film should be done after scanning

The marker and standard lanes

  • Bands beyond (i.e. to the right of) the last standard band won't be analyzed.
  • The standard file contains the band pattern of a marker lane from a master gel.

Project handling

  • the gel name is given when the image is scanned (or imported) and may contain alphanumeric characters.
  • it is advisable not to use spaces in filenames (especially when gels are created on a Macintosh, where spaces in filenames are quite popular - use underscores _ instead)
  • the project name used in image should be the same as used in the fpc database, to allow automatic transfer of data.
  • Image assumes that the directories for fpc are properly set up before a gel is transferred, whereas the scripts provided by Image will create directories for projects and gels used in image

Downloads

The Image software package is free.

The Image tutorial explains the software and what most parts of the program do.

If you are upgrading image, then PLEASE KEEP YOUR PREVIOUS VERSION. We primarily use Compaq alpha machines for software development and other platforms will be less well tested. There may also be other reasons why a new release won't work for you, so please be prepared to revert to a previous release while we fix any problems.

Download the latest version from our ftp server and then email image@sanger.ac.uk telling us who you are, where you're based and which platform (eg Solaris, Linux, etc) you are using Image on. Informing us of your name and email when you download the software helps us to keep track of the user community of Image. We can then inform you of bug fixes, new versions and noticeable improvements. Your personal data will be treated as strictly confidential.

If you have problems downloading the files, please contact us at image@sanger.ac.uk and we will do our best to help.

We havce compiled a summary of the new features/changes in im3.10.

Credits

im3 - graphical user interface
  • Fred Wobus (GTK-based user interface, editors, overall design)
  • Richard Durbin et. al. (original design of Acedb graphics library)
  • Simon Kelley (GTK adaptation of Acedb graphics library)
  • John Bradley (some Bitmap image manipuluation code from XV)
  • Darren Platt, Ken McDonald (additions and improvements)
analysis modules (amStepx)
  • Darren Platt (implementation of most C++ modules)
  • Jim Mullikin (revisions and improvements to various modules)
  • Steven Leonard (maintenance and adaptation of am-Modules)
  • Ken McDonald, Will FitzHugh (various contributions to analysis algorithms)
  • Original C-language conversion by Fred Wobus based on code by John Sulston et.al. and Richard Durbin
lane tracker - pathfinder
  • Jim Mullikin (design and implementation)
  • Steven Leonard (maintenance and additions)
image conversion program (im3conv)
  • John Cristy (Image-Magick Version 3.7.5)
  • Richard Durbin, Fred Wobus and others (ABI and other image conversion code)
* quick link - http://q.sanger.ac.uk/e5hdyd7m