Bayesian Inference for the Analysis of Sub-temporal-resolution Data

BIASD

BIASD allows you to analyze Markovian signal versus time series, such as those collected in single-molecule biophysics experiments, even when the kinetics of the underlying Markov chain are faster than the signal acquisition rate. The code here has been written in python for easy implementation, but unfortunately, the likelihood function is computationally expensive since it involves a numerical integral. Therefore, the likelihood function is also provided as C code and also in CUDA with python wrappers to use them with the rest of the code base.

Contents:

• Getting Started
  • Software Requirements
  • Adding BIASD to the Python Path
• Compiling the Likelihood
  • Background
  • How to Compile
  • Testing Speed
• BIASD Examples
  • Creating a new SMD file
  • Sample the posterior with MCMC
  • Laplace approximation and computing the predictive posterior
• BIASD GUI
  • Functionality

Code Documentation:

• Distributions
  • Some standard probability distributions
  • Convert between distributions
  • Distributions can be collected for priors or posteriors
  • Collections can be visualized
  • Finally, you can easily generate a few useful collections using
• Laplace
  • Laplace Approximation
• Likelihood
  • Switch the log-likelihood function
  • Inference-related functions
• MCMC
  • Markov chain Monte Carlo
  • Setup and Run MCMC
  • Analyze MCMC samples
• SMD
  • Work with SMD data
  • Add BIASD results to an SMD object
  • Read BIASD results from an SMD object into a useful format
  • Examples