Test the WNATO hypothesis

[matentzn]

One of the core components of the PhEval project is to understand how much of a difference organisms beyond Zebrafish and Mouse will make. We call this the, highly controversial, WNATO (We need all the organisms) hypothesis for VP (Variant Prioritisation). The hypothesis is that:

We can significantly increase the performance of VP when integrating gene and variant data of {speciesX}, where {speciesX} is something between Fission Yeast and Platypus.

The goal here is to generate enough experimental data to either support or reject support towards the expansion of Exomiser towards more than just Zebrafish and Mouse.

The preliminary plan here is this:

• Implement a PhEval experimental runner that can predict the right disease given a phenotypic profile (no other information)
• Create a new PhEval run config for studying the accuracy of various different KG components (KGX files and semantic similarity profiles)
• Write a paper that once and for all answers the question if Platypus should be included in Exomiser or not

[matentzn]

@souzadevinicius and @kcortes133 you will both be assigned certain sub-components of this task, and work together on reaching this great goal.

• @souzadevinicius you will work on semantic similarity, improving the ontologies, and probably their mappings. This includes a semsim-only method in the pheval.experimental runner which I will describe in a different ticket
• @kcortes133 you will probably focus on implementing the runner, running the experiments and writing the paper
• We will all work on the experimental design for the paper, but again, @kcortes133 you should lead the charge

This project is successful if we can either

1. Convince Melissa that integrating Platypus is detrimental to VP performance
2. Convince Damian that integrating Patypus is beneficial to VP performance.

Good luck.