Tasks

Network Characterization

Data

• Average degree for all the nodes in the full graphs.
  • Get full network at all levels, with and without small molecules:
    • Create all pathway networks
    • Implement function to retrieve all reactions part of pathways.
    • Implement function to get participants of a reaction.
    • Implement function to get complex components
• Calculate average, min, max degree from full networks.
• Calculate graph density. How does this help? The higher the density, the more edges are there compared to the number of nodes. If we have two graphs with the same number of nodes, then the higher the density then it means they are more connected, but less density they are more disconnected. If we get more nodes and the edges distribute across old and new nodes, then if there is a smaller density then there are less edges per node. It means they preserve the information, while making it more specific for certain nodes. If the density increases it means the number of edges increased more than the nodes, which means the nodes have higher degrees than before, then there are many more edges where the protein split into multiple proteoforms connecting to all of them.
• Calculate also densities for all pathways accross granularity levels.
• Calculate bridges for single pathways
• Articulation points for single pathways. It may not work for the full network, because high degree.
• Centrality.
• Study what is multiple test correction and why the size of the gene sets is important for this.
• How are gene interaction networks useful for biological purposes?

Visualization

• Write function that colors the different pathways in an interaction network
• Show the difference between the proteoform interaction network of a pathway and the gene or protein IN.
  • Choose a set of nodes to fix and then make the plot.
• Visualize bridges and articulation points of the network.
• Make a figure showing multiple pathways converted into one connected interaction network, highlighting with different colors the different pathways.
  • Function to merge graphs from multiple pathways.

Text

• Explain the levels of granularity
• Explain what we mean by proteoforms
  • Definition
  • How do we obtain them

General - Alternatives

• Read about drug targets

• Study Python

• Draft the extension proposal

• Prepare the figures of a pathway all levels

• Prepare figures of module overlap

• Prepare figures of percolation

• Show figures of the cases for the module pairs overlap

• Find papers related to drug targets and signaling, to see if that is what he meant

• Check module overlap variation in general:

  • Size
  • Connectedness

Detailed

• For the drug targets I need to check if the possible targets vary when we go from gene level to the protein or proteoform level. If it stays as target in both levels, then yes. Somehow the proteoform modules should say that the target.

We start with a list of drug targets, then we want to know if a protein makes a good target in a disease. Usually the targets are selected from databases at gene level. Then, what happens if we check specific information of the proteoforms, and try to verify if the drugs still make a good target for the disease.

A question arises here, how are proteoform modules built.

• Disease module variation.
  • Size variation
  • Bridge edges variation
  • Reduction of overlap size
  • Overlap stays and involves only modified proteins

Notes

For each topic, I need to check the big picture and then give particular examples. I need to find out the trend and then show examples.

About what:

• The size of the modules varies
• The number of bridge edges varies
• Cases where the overlap involves only modified proteins. Pointing out to important signaling mechanisms.

Library functions required

Get the data

• Read data from PheGenI
• Read data from Chebi
• Read data from Diseases of Jensen lab

Get numbers from the data

• Get sizes of modules
• Get size variation of modules across levels
• Get number of edges between two modules

Select particular examples

• Select examples of module size variation
• Select examples of connected ness variation between two modules
• Select examples of overlap reduction by a percentage

Make figures

• Make a scatter plot
• Plot two modules across levels