Tool to browse kmer sequences laid out in 2D approximating Huddinge distance and displaying selex enrichment. Current visualisation shows 10 local maxima laid out in polar coordinates such that the distance from origin is the mean log fold enrichment in the selex experiments. The other kmers that are enriched with z-score>2.58 are displayed in the (jittered) direction of their representative local maxima. The representative is the most enriched reached with steps of huddinge size one. The points are colored according to their distance to the representative. The circle is at the highest enrichment peak.
For running the software you need a tab separated file with header line, kmers on first column and annotation and enrichment number values on other columns. Example file is at tests/enriched_kmers_z99_HOXB13.tsv. You should get a session:
$ huddinge_tsne_browser -i tests/enriched_kmers_z99_HNF4A.tsv -t enriched_kmers_HNF4A 2>/dev/null
<script src="enriched_kmers_HNF4A.js" id="bd73b7e5-bd13-42c5-8a5c-c0bd4e3a1914"></script>
$ ls enriched_kmers_HNF4A*
enriched_kmers_HNF4A.html enriched_kmers_HNF4A.js
$
The file enriched_kmers_HNF4A.html contains a sample HTML file you can open in a web browser. To embed the plot in your own file, copy-paste the script tag from standard output where you want to have the plot.
The example plots can be viewed online at:
The system has bit of dependencies with holoviews >=1.9.2, jellyfish and jupyter covering most of them. Those are easiest to install from conda-forge and bioconda.
If you have bioconda installed (you should), then it's enough to say
conda install -c kpalin huddinge_tsne_browser pyhuddinge