CGView is a circular genome graphing utility. You can give it features encoded in XML files, and it generates a display of those features along a circular genome fragment.

Basic Structure

CGView accepts three different file formats, but we'll focus on XML here. The basic structure is this:

CG View settings
- Legend (optional)
  - LegendItem (text that you want in a legend)
- FeatureSlot 1 (this is a circular track)
  - Feature in FeatureSlot 1
  - Feature in FeatureSlot 1
- FeatureSlot 2
  - Feature in FeatureSlot 2
    ...
- FeatureSlot N
  - Feature in FeatureSlot N
    ...

In XML, this has the basic form:

Basic example from the CGView documentation

<?xml version="1.0" encoding="ISO-8859-1"?>
<cgview backboneRadius="160" sequenceLength="10000" height="600" width="600">

  <legend position="upper-right">     
    <legendItem text="Promoter" drawSwatch="true" swatchColor="green" />
    <legendItem text="Terminator" drawSwatch="true" swatchColor="red" />
  </legend>

  <featureSlot strand="direct">
    <feature color="green" decoration="clockwise-arrow" label="a promoter">
      <featureRange start="500" stop="900" />
    </feature>
  </featureSlot>

  <featureSlot strand="reverse">
    <feature color="red" decoration="arc" label="a terminator">
      <featureRange start="8500" stop="8800" />
    </feature>
  </featureSlot>
    
</cgview>

Basic Attributes

Here are some quick summaries of some useful attributes. Complete information is available from the CGView documentation.

featureSlot

Define a circular track in the graph.
strand direct or reverse affects the direction of arrows, and whether features are drawn on the inside or outside of the backbone circle.

feature

Define a feature in the graph.
color red, blue, green, purple are easy choices, but you can also use RGB values with something like rgb(0,225, 0).
decoration arc, clockwise-arrow, counterclockwise-arrow
proportionOfThickness real number between 0 and 1.0 changes the thickness of the feature.
radiusAdjustment real number between 0 and 1.0 adjusts the radius at which the feature is drawn. 0 draws the feature as close to the backbone circle as possible, 0.5 draws the feature in the middle of the featureSlot.

featureRange

Where feature begins and ends (and there can be more than one featureRange per feature).
start, stop integers between 0 and the sequence length. Each featureRange needs start and stop coordinates.
color, decoration, proportionOfThickness, radiusAdjustment have the same meaning as for a feature, and setting these values for a featureRange overrides the values in the encompassing feature.

Helper Scripts

Here's a quick Bash script that takes files in XML format representing featureSlot elements, and combines them into an XML file for CGView.

makeMap.sh

#!/bin/bash

cp header.xml cg3.xml
cp header_nogenelabels.xml cg3.xml
cat "$@" >> cg3.xml 
echo "</cgview>" >> cg3.xml 
cgview -i cg3.xml -o map.jpg -f jpg

This code depends on the following header file (which should be modified to fit the current use):

header.xml

<?xml version="1.0" encoding="ISO-8859-1"?>
<cgview backboneRadius="600" backboneColor="rgb(102,102,102)" backboneThickness="8" featureSlotSpacing="small" labelLineLength="100" labelP\
lacementQuality="good" labelLineThickness="2" rulerPadding="10" tickThickness="6" arrowheadLength="6" rulerFont="SansSerif, plain, 20" rule\
rFontColor="rgb(0,0,0)" labelFont="SansSerif, plain, 15" isLinear="true" minimumFeatureLength="0.2" sequenceLength="1255523" height="3000" \
width="3000" globalLabel="true" moveInnerLabelsToOuter="false" featureThickness="30" tickLength="15" useInnerLabels="true" shortTickColor="\
rgb(0,51,0)" longTickColor="rgb(0,51,0)" zeroTickColor="rgb(0,51,0)" showBorder="true" borderColor="black" backgroundColor="white" tickDens\
ity="0.166666666666667">




<!-- title -->
<legend position="lower-center" backgroundOpacity="0.8">
<legendItem textAlignment="center" font="SansSerif, plain, 60" text="IgH Locus (from chr14:106032563-107288085)" />
</legend>





<!-- details legend -->
<legend position="upper-left" font="SansSerif, plain, 20" backgroundOpacity="0.8">
<legendItem text="Accession: NC_000014" />
<legendItem text="Length: 1,255,523 bp" />
<legendItem text="Topology: linear" />
</legend>





<!-- legend -->
<legend position="upper-right" textAlignment="left" backgroundOpacity="0.8" font="SansSerif, plain, 20">
<legendItem text="CDS" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(0,0,153)" />
<legendItem text="tRNA" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(153,0,0)" />
<legendItem text="rRNA" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(153,0,153)" />
<legendItem text="Other" drawSwatch="true" swatchOpacity="0.5" swatchColor="rgb(51,51,51)" />
<legendItem text="Chr3 TLs" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(102,255,51)" />
<legendItem text="Chr5 TLs" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(102,51,255)" />
<legendItem text="Chr8 TLs" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(204,51,255)" />
<legendItem text="Chr16 TLs" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(255,51,102)" />
<legendItem text="Chr17 TLs" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(255,51,204)" />
<legendItem text="Chr18 TLs" drawSwatch="true" swatchOpacity="1.0" swatchColor="rgb(51,102,255)" />
</legend>

Using makeMap.sh

makeMap.sh FeatureA.xml FeatureB.xml

generates the graph as map.jpg, with Feature A and Feature B as separate tracks, where (for example) FeatureA.xml and FeatureB.xml are files of the form

FeatureA

<featureSlot showShading="false" strand="direct">
<feature color="green" decoration="arc" >
<featureRange start="1" stop="1184" />
<featureRange start="1198" stop="1367" />
<featureRange start="1416" stop="1607" />
<featureRange start="1614" stop="1853" />
</feature>
</featureSlot>

Note that the input files here aren't technically XML files themselves, since they don't begin with an XML declaration, but this is a minor technical point.

Graphing a histogram

Graphing a histogram in CGView using each position as a datapoint will probably take lots of computational time, with marginal benefits to resolution. Averaging over some region of positions is probably a more efficient way of getting a histogram. The following script takes a table of position counts (as a text file), and outputs an XML fragment of the counts averaged over regions of length window_size which can be used as input to makeMap.sh above.

The position count table must have the form
<column 1> <column 2, probably position> <int> <int> ... <int>
<column 1> <column 2, probably position> <int> <int> ... <int>
...
where each <int> is a count.

Script to average a count table

Average Count Table

#! /usr/bin/env python
import sys

try:
    import argparse
except:
    print 'Try typing "module load python" and then running this again.'
    sys.exit(1)


def main():
    parser = argparse.ArgumentParser(description='Averages read depth along chromosome positions from a number of samples.')
    parser.add_argument('count_table')

    args = parser.parse_args()

    window_table = []

    window_size = 250
    window_sum = 0
    line_count = 0

    with open(args.count_table, 'r') as count_table:
        for line in count_table:
    
            fields = line.split()
            count_fields = [int(x) for x in fields[2:]]
            window_sum += sum(count_fields)

            line_count += 1
            if (line_count % window_size) == 0:
                window_table.append([line_count-window_size+1, line_count, window_sum])
                window_sum = 0

        if (line_count % window_size) != 0:
            window_table.append([line_count - (line_count % window_size) + 1, line_count, window_sum])

    max_count = 0
    for window in window_table:
        if max_count < window[2]:
            max_count = window[2]

    for window in window_table:
        window[2] = float(window[2])/max_count

    print '<featureSlot strand="direct">'
    print '<feature color="purple" decoration="arc">'
    for window in window_table:
        print '<featureRange start="{}" stop="{}" proportionOfThickness="{}" />'.format(window[0], window[1], window[2])
    print '</feature>'
    print '</featureSlot>'

if __name__ == '__main__':
    main()

Space shortcuts

Page tree