No packages match

curatedPCaData - Curated Prostate Cancer Data

The package curatedPCaData offers a selection of annotated prostate cancer datasets featuring multiple omics, manually curated metadata, and derived downstream variables. The studies are offered as MultiAssayExperiment (MAE) objects via ExperimentHub, and comprise of clinical characteristics tied to gene expression, copy number alteration and somatic mutation data. Further, downstream features computed from these multi-omics data are offered. Multiple vignettes help grasp characteristics of the various studies and provide example exploratory and meta-analysis of leveraging the multiple studies provided here-in.

Last updated

experimenthubexperimentdataprostatecancerdatacancerdatahomo_sapiens_datamicroarraydatarnaseqdataexpressiondatacopynumbervariationdatasomaticgeoreproducibleresearchcancercancer-datacancer-genomicscancer-researchdatapkg

5.82 score 11 stars 12 scripts

ePCR - Ensemble Penalized Cox Regression for Survival Prediction

The top-performing ensemble-based Penalized Cox Regression (ePCR) framework developed during the DREAM 9.5 mCRPC Prostate Cancer Challenge <https://www.synapse.org/ProstateCancerChallenge> presented in Guinney J, Wang T, Laajala TD, et al. (2017) <doi:10.1016/S1470-2045(16)30560-5> is provided here-in, together with the corresponding follow-up work. While initially aimed at modeling the most advanced stage of prostate cancer, metastatic Castration-Resistant Prostate Cancer (mCRPC), the modeling framework has subsequently been extended to cover also the non-metastatic form of advanced prostate cancer (CRPC). Readily fitted ensemble-based model S4-objects are provided, and a simulated example dataset based on a real-life cohort is provided from the Turku University Hospital, to illustrate the use of the package. Functionality of the ePCR methodology relies on constructing ensembles of strata in patient cohorts and averaging over them, with each ensemble member consisting of a highly optimized penalized/regularized Cox regression model. Various cross-validation and other modeling schema are provided for constructing novel model objects.

Last updated

5.04 score 22 scripts 270 downloads

oscar - Optimal Subset Cardinality Regression (OSCAR) Models Using the L0-Pseudonorm

Optimal Subset Cardinality Regression (OSCAR) models offer regularized linear regression using the L0-pseudonorm, conventionally known as the number of non-zero coefficients. The package estimates an optimal subset of features using the L0-penalization via cross-validation, bootstrapping and visual diagnostics. Effective Fortran implementations are offered along the package for finding optima for the DC-decomposition, which is used for transforming the discrete L0-regularized optimization problem into a continuous non-convex optimization task. These optimization modules include DBDC ('Double Bundle method for nonsmooth DC optimization' as described in Joki et al. (2018) <doi:10.1137/16M1115733>) and LMBM ('Limited Memory Bundle Method for large-scale nonsmooth optimization' as in Haarala et al. (2004) <doi:10.1080/10556780410001689225>). The OSCAR models are comprehensively exemplified in Halkola et al. (2023) <doi:10.1371/journal.pcbi.1010333>). Multiple regression model families are supported: Cox, logistic, and Gaussian.

Last updated

fortranpenalized-regressionregression

4.38 score 2 stars 12 scripts 296 downloads

hamlet - Hierarchical Optimal Matching and Machine Learning Toolbox

Various functions and algorithms are provided here for solving optimal matching tasks in the context of preclinical cancer studies. Further, various helper and plotting functions are provided for unsupervised and supervised machine learning as well as longitudinal mixed-effects modeling of tumor growth response patterns.

Last updated

4.19 score 2 dependents 26 scripts 315 downloads