Data

Tools

Indexes

Applications

Experiments

Open Source Science

https://github.com/broadinstitute/getzlab-lnp

Log-normal-Poisson regression model from Hess et al. 2019

https://github.com/broadinstitute/getzlab-lnp

Science Score: 36.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ○
    CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ○
    .zenodo.json file
  • ✓
    DOI references
    Found 2 DOI reference(s) in README
  • ○
    Academic publication links
  • ✓
    Committers with academic emails
    1 of 1 committers (100.0%) from academic institutions
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (13.3%) to scientific vocabulary
Last synced: 10 months ago · JSON representation

Repository

Log-normal-Poisson regression model from Hess et al. 2019

Basic Info
Statistics
  • Stars: 5
  • Watchers: 3
  • Forks: 2
  • Open Issues: 2
  • Releases: 0
Created about 7 years ago · Last pushed almost 7 years ago
Metadata Files
Readme

README.md

Log-normal-Poisson regression model

Code to run the log-normal-Poisson regression model from Hess et al. 2019, "Passenger Hotspot Mutations in Cancer" (https://doi.org/10.1016/j.ccell.2019.08.002)

What's in this repo?

This repository contains MATLAB functions for running a Bayesian log-normal-Poisson (LNP) regression. We include code to both sample from the posterior distribution of the model parameters via MCMC, and also compute statistical significance of counts, given a set of posterior samples (i.e., compute posterior predictive p-values).

The aforementioned functionality is suitable for modeling generic count data, and is not specific to modeling somatic mutation counts. For that purpose, we also include a function to process the specific somatic mutation calls analyzed in Hess et al.

Included tools

  • To run the regression on raw count/covariate data, use src/regr/pois_LN_reg.m (see demo.m, or the inline example in this readme)
    • This tool generates samples from the posterior distribution of the LNP parameters (μ, τ, β); it does not identify significant outlying counts.
    • To compute significance (i.e., LNP posterior predictive p-values), use src/regr/regr_post_pred.m
  • To run the model on the somatic mutation calls analyzed in the manuscript, use the wrapper script src/pois_LN_reg_wrapper.m. Currently, it is only intended for processing calls in a format specific to this manuscript, output by the analysis notebooks used to generate the publication. We are planning to release a user-friendly wrapper for running on generic .maf files soon.

To run any code, start MATLAB (any version R2014b or newer should work) in the root directory of this repo --- this is necessary for startup.m to properly add dependencies to the MATLAB path. Tested to work only under 64 bit Linux; other architectures may work after recompiling C/C++ .mex files.

Regression demo

Here is a simple demo of running the regression model on simulated counts, sans covariates.

First, we generate 50,000 random samples from a log-normal-Poisson distribution with μ = -3, σ = 0.9:

MATLAB x = poissrnd(exp(-3 + 0.9*randn(50000, 1))); We thus expect samples drawn from the posterior distribution p(μ,σ|x) will be centered around (-3, 0.9).

Next, we set some basic parameters for running the MCMC:

``` MATLAB P = []; P.niter = 2000; % total iterations P.burnin = 50; % burn-in P.m0 = log(mean(x)); % initial guess for mu

% normal-gamma hyperparameters P.mumu = -3; P.taua = 10; P.taub = 0.2; ```

Actually run the MCMC:

MATLAB [~, ~, mu, tau] = pois_LN_reg(x, zeros(size(x)), P);

2,000 samples from μ and σ will be placed into mu and tau, respectively. Note that τ = 1/σ^2.

Finally, we plot samples from the posterior:

``` MATLAB figure(1); clf hold on scatter(mu(P.burnin:end), 1./sqrt(tau(1, P.burnin:end)), 'marker', '.', ... 'markeredgealpha', 0.6, 'markeredgecolor', 'k') scatter(-3, 0.9, 70, 'marker', 'x', 'markeredgecolor', 'm', 'linewidth', 2)

title('MCMC draws from LNP posterior p(\mu, \sigma|x)')

xlabel('\mu') ylabel('\sigma')

ax = gca; ax.Box = 'on';

ax.XLim = [-3.15 -2.85]; ax.XTick = -3.15:0.05:-2.70;

ax.YLim = [0.75 1.05]; ax.YTick = 0.75:0.05:1.05;

grid on ```

MCMC samples from posterior

confirming that samples from the posterior distribution look as we would expect.

Owner

  • Name: Broad Institute
  • Login: broadinstitute
  • Kind: organization
  • Location: Cambridge, MA

Broad Institute of MIT and Harvard

GitHub Events

Total
Last Year

Committers

Last synced: about 1 year ago

All Time
  • Total Commits: 4
  • Total Committers: 1
  • Avg Commits per committer: 4.0
  • Development Distribution Score (DDS): 0.0
Past Year
  • Commits: 0
  • Committers: 0
  • Avg Commits per committer: 0.0
  • Development Distribution Score (DDS): 0.0
Top Committers
Name Email Commits
Julian Hess j****s@b****g 4
Committer Domains (Top 20 + Academic)
broadinstitute.org: 1

Issues and Pull Requests

Last synced: about 1 year ago

All Time
  • Total issues: 2
  • Total pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Total issue authors: 2
  • Total pull request authors: 0
  • Average comments per issue: 1.5
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
Past Year
  • Issues: 0
  • Pull requests: 0
  • Average time to close issues: N/A
  • Average time to close pull requests: N/A
  • Issue authors: 0
  • Pull request authors: 0
  • Average comments per issue: 0
  • Average comments per pull request: 0
  • Merged pull requests: 0
  • Bot issues: 0
  • Bot pull requests: 0
View more stats
Top Authors
Issue Authors
  • gxiefanfan123 (1)
  • a00101 (1)
Pull Request Authors
Top Labels
Issue Labels
Pull Request Labels