Journal: Database: The Journal of Biological Databases and Curation

Article Title: An entropy-reducing data representation approach for bioinformatic data

doi: 10.1093/database/bay029

Figure Lengend Snippet: 2D sub-spectra for all EST sequences in the dataset. Each point in this plot corresponds to one EST sub-spectrum, but often corresponds to multiple ESTs (which have the same sub-spectrum). Whereas the full EST spectrum has 10 elements, the sub-spectra depicted here have 2 elements.

Article Snippet: Example 1: Alignment of sequences against multiple references In this example we aligned a dataset of 203 333 cattle expressed sequence tags (ESTs) (AgResearch, Genesis and Primary Industry Victoria Bovine EST Project 2006, unpublished data) to a panel of 10 sheep and cattle genomes and transcriptomes using BLAST, in order to investigate the possibility of non-semantic annotation of these sequences, using alignments against multiple reference genome and transcriptome assemblies to embed the sequence collection in a metric representation space in which distances between sequence representations are related to biological distance; and investigate the possibility of using the dual embedding of the panel of genomes and transcriptomes to investigate and visualise the relationships between them, via a suitable metric on the dual representation space.

Techniques: