sadcricket co. ltd.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
Chr20FuncIntx.tsv
Chr20GOslimData.tsv
Chr20GWAStraits.tsv
Chr20GeneData.tsv
README-DATA
README.md
circosStarter.py
genomePlotBasic.R
genomePlotDemo.R
genomePlotFunctions.R
genomePlotIntermediate.R
hill_halo.png
hill_halo_demo.png
linearStarter.py
mart_export.txt
ppi_0.png
ppi_community.png
ppi_plot-v1.ipynb
ppi_plot.ipynb
prepCh20Data.py
prepareGenomeData.R
sadCricket.png
setUpCh.py
svgCols.txt
test.svg

README.md

bcbbiohacks2018

sadcricket co. ltd.

Graph library: networkx Community detection: girvan_newman Community layout: networkx.spring_layout Plotting the hills: 3d Gaussian All plots: matplotlib.pyplot

PPI layer:

  • layout of communities with spring, st that more connected communities are closer to each other
  • Size of communities along x-y plane (i.e. sigma of Gaussian) proportional to number of genes in community
  • FUTURE: mu of Gaussian in communities can be shifted according to centrality of nodes in the community
  • FUTURE: Height of communities along z-axis should be inversely proportional to number of GO / number of genes; i.e. communities with many similar GO term mappings have higher peaks
    • So then expect thicker PPI-GO lines for higher peaks

GO layer:

  • Each dot is a GO term, with colors for namespace:
    • Red = P
    • Blue = C
    • Yellow = F
  • FUTURE: size of dots is number of proteins mapped to it
  • FUTURE: GO terms are arranged according to tree hierarchy, so closer to centre of circle means it’s lower in the tree i.e. more specific; GO terms close to each other are closer in the tree

GWAS layer:

  • Have a server ready, input a GWAS feature i.e. “diabetes-associated”
  • FUTURE: highlight proteins (points on surface of the hills) with this GWAS feature, the community it belongs to, and the GO terms it maps to

~7-10 dimensions represented in a 3D space