vatj / gpmap_integer_polyomino

Code repository for the analysis of the Genotype-Phenotype Map Properties of the Integer Polyomino Model.

Genotype-Phenotype Map Integer Polyomino

Code repository for the analysis of the Genotype-Phenotype Map Properties of the Integer Polyomino Model.

Overview

Custom GP-map analysis tools for the Integer Polyomino Model are coded in C++. These tools rely on a core algorithm to build polyominoes from lists of integer hosted in this repository. The python API uses pybind11 to expose the underlying C++ to the python interpreters. A minimal documentation of the python API is provided using Sphinx and hosted on the Theory of Condensed Matter Group servers. Additionally, example jupyter notebooks are made available to get started with the python API.

Installation

This module can be installed as any python package :

pip install git+https://github.com/vatj/gpmap_integer_polyomino.git

Note that compiling requires a relatively recent C++ compiler such as g++7 as well as openmp libraries to enable parallel computing

Quickstart

Two main quickstart jupyter notebooks are provided to examplify the basic functionality of the repository. Other examples notebooks are also present which can be used for more advance use cases.

Assembly

This notebook contains a minimal example of how to use the polyomino assembly API. It is intended as a tool to map a list of integer to a polyomino, i.e. a genotype to its corresponding phenotype. In practice the function only returns an ID which corresponds to an entry in a phenotype_table file. Each entry is a matrix of boolean value. A simple plot function is provided to map this matrix onto a polyomino shape.

GP-Map

This notebook contains an example of how to generate the the analysis of a integer polyomino genotype-phenotype map. All genotype contains 2 genes and at most 7 different labels (1 neutral and 3 interacting pairs). The output is written to 2 different files, one for genome-level metrics and one for phenotype-level metrics. These files can be loaded in panda dataframes for easy analysis and plotting.

Directory Layout

• docs/ (integration with sphinx)
• libs/ (git submodule dependencies)
• notebooks/ (working examples)
• src/
  • integer_polyomino/
    • assembly/ (Mostly pybind API for custom polyomino_core features)
    • gpmap/ (C++ implementation of the gpmap analysis algorithms + pybind API definition)
    • scripts/ (python utility file such as plotting functions)
• tests/

gpmap_integer_polyomino's Issues

Consistent duplicate generation

What

Duplicates should be taken from the same list that generated the 3 tile space

Why

Removes confusion over parentless duplicates and childless 3 tiles.

How

Save 3 tile list to file, read file, duplicate based on given list
(do we want to strip isomorphic duplicates, it is meaningful that they are twice as common...)

Clarify "quickstart"

It is confusing to install and use this repo from scratch.

• clarify state system requirements (CMake etc.)
• improve usage of quickstart files
  • remove clutter (privateSainsburyPresentation etc.)
  • comment more on user interface side
  • command line option?
• docs and tests folders are basically nonfunctional

Genotype/Assembly graph feature identification

What

ID genotype features like

• free faces
• total number of interactions
• assembly graph features
  • branching points
  • cycles
  • SIFs

Why

Better understand phenotype space in general, and specifically how duplication effects different subunits.

How

reconstruct info based on the reference genotype present in set metric file

Packing error

Does graph_topo need to be additionally added to local directory?

If we make a pypi package, it should be 1 click installable, not require many steps.