This repo contains the unformatted tex document for the publication "Detection of Markers for Discrete Phenotypes". The paper was published and is available at
A video of the presentation at CSBio21 is available here
If you have questions, want to discuss ideas or report errors and typos in the manuscript please open an issue or contact
The repo also contains a Python CLI tool, called biomarkers, that can be used to compute the markers for a given Boolean network and phenotype components.
To install the master branch use:
pip3 install pip --upgrade
pip3 install --force-reinstall git+https://github.com/hklarner/detection_of_markers_for_discrete_phenotypes
To install a tagged version use the @:
pip3 install pip --upgrade
pip3 install git+https://github.com/hklarner/detection_of_markers_for_discrete_phenotypes@1.0.0
If the installation was successful you should be able to see the help menu for the biomarkers tool:
$ biomarkers -h
Usage: biomarkers [OPTIONS] COMMAND1 [ARGS]... [COMMAND2 [ARGS]...]...
Options:
-v, --version Display version.
-h, --help Show this message and exit.
Commands:
control-create Create a control file.
control-export Export a control file.
json-info Reads json file and prints summary.
markers-export Exports markers as CSV.
markers-factorize Factorizes a marker set.
markers-count Creates a marker count graph.
markers-info Prints info about a markers file.
markers-validate Validates a marker set.
problem-create Creates a marker detection problem.
problem-info Displays info about a problem file.
problem-solve Solves a marker detection problem.
repo Access to the pyboolnet repository.
steady-state-correlation Computes the blocks of the steady state correlation.
steady-state-matrix Prints the steady state matrix.
To see the help text and available options of a command call the command with the option -h.
E.g., to see the help text for the command markers-create call biomarks markers-create -h.
The computation of the markers is a two-step process.
First, define a problem file by specifying a Boolean network and the phenotype subspace using the command problem-create:.
$ biomarkers problem-create --problem emt_problem.json --bnet selvaggio_emt --phenotype AJ_b1=0,AJ_b2=0
created json file: fname=emt_problem.json
n_steady_states: 1452
component_names: {0: 'AJ_b1', 1: 'AJ_b2'}
phenotype_subspace: {0: 0, 1: 0}
phenotype_index phenotype n_steady_states
0 0 not {0: 0, 1: 0} 184
1 1 {0: 0, 1: 0} 1268
Instead of specifying a phenotype subspace you may specify the phenotype components:
$ biomarkers problem-create --problem emt_problem.json --bnet selvaggio_emt --phenotype AJ_b1,AJ_b2
created json file: fname=emt_problem.json
n_steady_states: 1452
component_names: {0: 'AJ_b1', 1: 'AJ_b2'}
phenotype_components: [0, 1]
phenotype_index phenotype n_steady_states
0 0 {0: 0, 1: 0} 1268
1 1 {0: 1, 1: 1} 148
2 2 {0: 1, 1: 0} 36
The command creates a problem file in json format.
To compute the markers for a problem file use the command problem-solve:
$ biomarkers problem-solve --problem emt_problem.json --forbidden AJ_b1,AJ_b2 --marker-size-max 5 --markers emt_markers.json
created tmp_program.asp
clingo --models=0 --opt-mode=optN --enum-mode=domRec --heuristic=Domain --dom-mod=5,16
cpu time: 0:02:09.400305
n_marker_sets=116
first 3 marker sets: [[4, 54], [4, 53], [4, 44]]
created json file: fname=emt_markers.json
The command creates a markers file in json format.
To export a marker set in csv format use the command biomarkers markers-export:
$ biomarkers markers-export --markers emt_markers.json --csv emt_markers.csv
index name count likelihood
2 AKT 8 0.07
4 BCat_AJ 4 0.03
9 ECM 8 0.07
10 ECad 29 0.25
...
created emt_markers_counts.pdf
created emt_markers_counts.tex
The command markers-count counts how often network components appear as markers.
The option --pdf maps the counts onto an interaction graph and --tex export the data as a tex table.
$ biomarkers markers-count -m emt_markers.json --pdf emt_markers_counts.pdf --tex emt_markers_counts.tex
To factorize a marker set use the command biomarkers markers-factorize and specify the markers file.
The option --pdf creates, for each row in the table, a plot of the interaction graph with the marker subsets that make up the factorization highlighted:
$ biomarkers markers-factorize --markers emt_markers.json --pdf emt_markers_factorization
n_components n_markers factorization n_factors_used n_factors_available
2 16 S(4,11,12,31) * S(10,44,53,54) 2 2
3 76 S(10,44,53,54) * L(n=2,k=19) 2 2
4 24 S(42,43) * S(10,44,53,54) * {{24,13},{25,13},{24,14}} 3 3
created emt_markers_factorization_n2.pdf
created emt_markers_factorization_n3.pdf
created emt_markers_factorization_n4.pdf