Features
Main Frame of Meander Visualization
1) A Hilbert representation of the zoomed area (zoom level 3) of a chromosome.
2) A linear representation of the zoomed part of the chromosome.
3) A Hilbert representation of the whole chromosome.
4) The selected area of a chromosome.
5) Drop-down menu to adjust the zoom level.
6) A linear representation of the whole chromosome. The blue area indicates the part of the chromosome that users zoomed in.
7) Histogram for the read-depth coverage of the sample and the reference individually and dynamic filter to place threshold on the coverage.
8) Histogram for the log2 ratio between the reference and the sample and dynamic filtering.
9) Search box to search by position.
10) Drop-down box to switch visualization between sample, reference and ratio.
11) Overlay pre-calculated predicted variations.
12) Threshold according to the number of pair-ends that overlap and form a cluster.
Pair-End Information
Combination of read-depth and pair-end information as double evidence for insertions and deletions between two Arabidopsis Thaliana strains. Strain ICE153: Central Asia, Sequencing depth 21X, Latitude 41,45, Longitude 70,05. Strain ICE97: Southern Italy, Sequencing depth 19X, Latitude 41,62, Longitude 12,87.
A) A Hilbert curve representation the log2 ratio between the two strains. Blue signals indicate possible insertions whereas yellow signals indicate possible deletions.
B) A Hilbert curve showing the pair-end mapping of the zoomed region.
C) A linear histogram showing both pair-ends and read-pair information. A very intuitive representation that shows double evidence about insertions and deletions as both PE and RD information is combines. The yellow signal drop indicates a deletion based on RD information. Additional information is given when the pair-ends map to corresponding positions of the deletion. Similarly the blue signal gains indicate an insertion and the pair-end mappings correspond to the relevant positions.
D) A Hilbert representation showing both the PE and the RD information of the whole chromosome. Only the pair-ends that belong to the selected area and map somewhere else in the chromosome outside the zoomed area are shown.
E) A linear representation of the pair-ends that belong to the selected area and map somewhere else in the genome outside the zoomed area.
Overlaying variations
Here we see an example of how Meander is able to overlay variations as they have been predicted by other external tools.
The specific example shows chromosome 22 without any deletions overlaid (left picture). Wee see that some areas with low coverage (light gray) might indicate some deletions and similarly, areas with high coverage might indicate some insertions.
The picture on the right shows how predicted deletions from SNV-SEQ software are overlaid on top of the Hilbert curve.
Meander is indeed a useful tool to observe visual patterns as the deletions that were predicted by SNV-SEQ overlap with the light gray areas that previously indicated possible deletions.
Sample vs Control
Meander can be used to compare the differences between a control and a sample genome. This could be for example genomes from a patient and a healthy person accordingly.
Some patterns can be observed visually such as areas with lighter signals (indicating deletions) or areas with higher signals comparing to the reference genome (indicating insertions).
Comparing two samples
Similarly to microarrays, Meander offers two color-schemes (red-green and blue-yellow) to compare reference and sample genomes.
To compare there genomes we use the log(sample/reference) value. When the sample is higher than the control (indicating duplications) the log value>0 and the pixel is painted green. similarly, when the sample is lower than the control (indicating a deletion) the pixel is painted red.
The picture on the right shows both a linear and a Hilbert representation of chromosome 1 using the log ratio.
Multiple sample comparison
Meander is able to compare up to four different samples against a common reference. The panels on the left are selectable and only one of them can be visualized at a time on the main screen.
Visualization of SVs that are supported both by read-pair and pair-end information
As only pair-end or only read-depth information is not sufficient evidence for a real structural variation, Meander is able to highlight those areas that are supported by double evidence.
Those areas are calculated automatically and highlighted with a circle
Inter-chromosomal variations
A line of adjusted length is wrapped around the main panel of the application indicating the whole genome. The genome is split onto as many segments as the organism that it refers to.
the length of the segments is adjusted according to the real chromosome lengths.
When an inter-chropmosomal variation occurs, the pair-end that corresponds to another chromosome is mapped to a specific position.
The chromosome that is loaded on the main screen is always highlighted in red.
Interactive Thresholds
The figure shows coverage of chromosome 1 reference genome.
A cutoff of 100 was set to remove the pixels with coverage less than 100 and therefore more white areas are observed in our visualization.
In the case of ratio where values range between -n<0<n both an upper and a lower threshold can be applied.
This is an important feature because just by looking the Hilbert coloring it is difficult to distinguish between a very high pick from the noise.
Lets say for example that one read has coverage 1000 and another read has coverage of 300 while most of the values range between 0-300. Both the coverage of the 1000 and the coverage of 300 will have the same color code indicating the intensity. A filter then, is necessary to distinguish the aforementioned.
Similar threshold can be applied also in:
