figshare
Browse
poster.pdf (2.85 MB)

Sequence Logo Redesign

Download (0 kB)
poster
posted on 2013-09-28, 09:26 authored by Eamonn MaguireEamonn Maguire, Philippe Rocca-Serra

My sequence logo redesign for Biovis 2013. See link below for original sequence logo design and the spec for the competition.

 

In this work, we investigate how the sequence logo could be improved from both the perspective of a bioinformatician and computer scientist and a biologist. We identify the major changes invoked alongside the effect we wished to invoke.

In this design, we mixed a combination of the existing representation to decrease the time to learn and new glyph-based techniques to add additional information.

We have kept colors in place as a mechanism of identifying the classes of amino acids (aliphatic, hydroxyl/sulphur containing, cyclic, aromatic, basic and acidic), where for example, blue corresponds to the basic group. The consistency between the old and new figure in using the same colours is seen as a way of reducing the time to learn this new technique. Letters are kept in the encoding since they are well known by the biologists, however we remove those letters that would be too small to show to reduce visual clutter. Rectangles are used to fill space relative to the presence of an amino acid instead of the stacked letter approach used in the origin figure. We believe that this removes problems with the original representation where more dense letters, e.g H or A would look bigger compared to an I even though the actual size was the same.

Glyph visualisation techniques can be used to depict hydrophobicity of a region and general side-chain charge of the amino acids in the group. In our design we show these changes when there is a change between groups, see positions 4, 7 and 24, 27 in the figure showing changes in the liking of water. There are also some examples in changes in dominance of side-chain electrical charge at positions 24 and 29. We employ a technique published by Brandes et al. in 2013 termed GestaltLines (DOI 10.1111/cgf.12104) to give a simplified indication of region conservation. Through this technique, it is simple to see conserved regions without the need to parse any additional information in the image. The salient information, which in this case is conservation in regions of the protein are instantly accessible. We also use opacity to fade out highly unconserved regions to ease visual search. Moreover, each region is better separated visually for comparison purposes through the use of a light-grey background alternating with white.

 

History