Software of the CSB group
- Lux: A probabilistic generative model for quantification of DNA modifications
- AdaptiveGP: An implementation of a fully nonstationary, heteroscedastic Gaussian process for Matlab.
- MixChIP: A probabilistic method for cell type specific protein-DNA binding analysis
- BinDNase: A discriminative approach for transcription factor binding prediction using DNase I hypersensitivity data
- DyNB: A probabilistic method to analyze RNA-seq time series data.
- Sorad: A nonparametric probabilistic method to infer and analyze dynamic signaling pathways.
- LIGAP: A probabilistic method to identify condition/lineage specific time-course profiles.
A probabilistic model and MCMC sampler for reconstructing cell type
specific gene expression profiles from measurements of heterogeneous
tissues. No manual purification, such as laser capture microdissection,
is needed, but the user must provide (prior) information about cell
type proportions in each sample. Samples can also be categorized into
experimental conditions -- the model then makes predictions on
expression profiles across both cell types and experimental conditions.
A web tool
is also available. DSection is also included in an R package called CellMix.
A probabilistic protein-protein interaction guided method for
competitive transcription factor binding prediction. A
Matlab implementation of our fully probabilistic protein-protein
competitive transcription factor binding prediction method: MultiTF-PPI
page. A more
friendly user interface is under development.
Transcription factor binding prediction with multiple data fusion. A
Matlab implementation of
our transcription factor binding prediction method that can incorporate
multiple genome-level data sources: ProbTF
software page. We
have also developed a web tool that is publicly available at www.probtf.org.
- GPODE: Learning
the structure of gene regulatory networks using non-parametric
molecular kinetics. A
of Matlab functions that implement our gene regulatory network
inference method. The method can use time-series and steady state gene
expression (and protein) measurements and makes Bayesian inference for
the network structure using Gaussian process based non-parametric
molecular kinetics. GPODE
- DBNRJMCMC (*coming soon): Structural learning of dynamic Bayesian
networks. A set
of Matlab functions
that implement our RJMCMC and approximative MCMC DBN structure learning
from time series and/or steady state measurements will be added
shortly. In the meantime, please contact us directly via email.
- Robust periodicity detection:
- A set of Matlab functions
that implement our robust rank-based periodicity detection method can
be downloaded from the supplementary web
page and from here.
- A set of Matlab functions that implement our robust regression-based
periodicity detection method can be downloaded from the supplementary web
- GeneCycle R package at CRAN.
- Toolbox for Boolean and probabilistic Boolean networks. A
comprehensive toolbox to
work with Boolean networks and probabilistic Boolean networks can be
downloaded from PBN web page
maintained by Ilya