Back to Home

Software

• space (Sparse PArtial Correlation Estimation): space is an R package for estimation and identification of non-zero partial correlations via sparse regression techniques. It can be downloaded from http://cran.r-project.org/;  or by clicking space. It is useful in construction of large networks. For more details, see the paper: Peng, Wang, Zhou and Zhu (2009). Partial Correlation Estimation by Joint Sparse Regression Models, Journal of the American Statistical Association, Vol. 104, No. 486, 735-746 [technical report: pdf; arXiv:0811.4463]

• fpca (Functional Principal Component Analysis): fpca is an R package for estimation eigen-values and eigen-functions of the convariance kernel (fpca) via sparsely observed functional data. It can be downloaded from http://cran.r-project.org/;  or by clicking fpca. It is useful in longitudinal studies. For more details, see the paper: Peng and Paul (2009). A geometric approach to maximum likelihood estimation of the functional principal components from sparse longitudinal data, Journal of Computational and Graphical Statistics, 18(4): 995 - 1015 [pdf]  (technical report: arXiv:0710.5343v1); 1-29-09: updated R package fpca

• remMap (REgularized Multivariate regression for identifying MAster Predictors): remMap is an  R package for fitting multivariate regression models under high-dimension-low-sample-size setting. It can be downloaded from http://cran.r-project.org/; or by clicking remMap. It is useful in construction of networks by using two types of high dimensional data, e.g., CGH array and expression array. For more details, see the paper: Peng,  Zhu, Bergamaschi, Han, Noh, Pollack and Wang (2010) Regularized Multivariate Regression for Identifying Master Predictors with Application to Integrative Genomics Study of Breast Cancer, The Annals of Applied Statistics, 4 (1): 53-77 [technical report: pdf; arXiv:0812.3671v1].

• dynamics: dynamics is an R package for fitting autonomous dynamical systems using spline basis (B-splines or cubic polynomial splines). It can be downloaded by clicking dynamics. dynamics is useful for fitting the underlying common (autonomous) systems nonparametrically for a group of random curves when only a snapshot of each sample curve is observed. For more details, see the paper: Paul, Peng and Burman (2011) Semiparametric modeling of autonomous nonlinear dynamical systems with applications, The Annals of Applied Statistics, 5(3): 2078-2108 [technical report: pdf, arXiv:0906.3501v1]

• BINCO (Bootstrap Inference for Network COnstruction): BINCO is an R package to calculate the optimal threshold of selection frequencies for variable/feature selection through directly controlling the false discovery rate. It can be downloaded from http://cran.r-project.org/. It is useful for determining the amount of regularization in high-dimension regularization problems, particularly, unsupervised learning such as network structure learning. For more details, see the paper : Li, Hsu, Peng and Wang (2013). Bootstrap inference for network construction, The Annals of Applied Statistics, 7(1): 391-417. [Full Text].

• dagbag: dagbag is an R package to conduct DAG learning by hill climbing algorithm and bootstrap aggregating. It can be downloaded from https://github.com/jie108/dagbag. dagbag is useful for high-dimensional directed acyclic graph (DAG) model learning. For more details, see the paper: R. Wang and J. Peng. Learning directed acyclic graphs via bootstrap aggregating (2014). [arXiv:1406.2098]

• spaceMap: spaceMap is an integrative –omics analysis pipeline for constructing networks and conducting network analysis. It is maintained on https://topherconley.github.io/spacemap/. spaceMap is useful for high-dimensional conditional graphical model learning. For more details, see the paper: Conley C., Umut Ozbek, Wang P., and J. Peng. Characterizing functional consequences of DNA copy number alterations in breast and ovarian tumors by spaceMap (2018). Journal of Genetics and Genomics, 45(7): 361–371. [bioRxiv: 10.1101/248229]

• PLNet (Poisson Lognormal Network): PLNet is an R package to infer graphical models and build networks based on RNA-Seq data. It can be downloaded from https://github.com/jie108/PLNet/. For more details, see the paper: Choi Y., Coram M., J. Peng, and Tang H. A Poisson Log-Normal model for constructing gene covariation network using RNA-seq data (2017). Journal of Computational Biology, 24(7): 721-731. [pdf]

• TSNet (Tumor Specific Network): TSNet is an R package for constructing tumor-specific networks based on samples with tumor purity heterogeneity. It can be downloaded from https://github.com/jie108/TSNet. For more details, see the paper: Francesca Petralia, Li Wang, J. Peng, Arthur Yan, Jun Zhu, and Pei Wang. A new method for constructing tumor specific gene co-expression networks based on samples with tumor purity heterogeneity. Bioinformatics, 34(13): i528–i536, 2018.
  [html]

• loggle (LOcal Group Graphical Lasso Estimation): loggle is an R package for constructing time-varying graphical models. It can be downloaded from http://cran.r-project.org/. For more details, see the paper: Jilei Yang and J. Peng. Estimating time-varying graphical models (2018). [arxiv:804.03811].