Software

This software analyses multiple data arising from clinical oncology (routine care and clinical trials). This data can be either static (e.g., baseline features), from clinical, biological, molecular (e.g., transcriptomic or mutation data) or
longitudinal (multiple time points per individual): tumor kinetics, biomarkers. The second type is modeled using the framework of nonlinear mixed-effects modeling. All features are then analyzed using data science techniques (preprocess, feature selection, machine learning algorithms), in order to predict survival outcome.

The package compoEDA aims to provide a comprehensive exploratory analysis of data from clinical studies in oncology. These studies commonly investigate biological markers able to reveal and distinguish different tumor profiles, in order to early adapt the therapeutic strategy for patients.

The objective of this software is to provide a simplified tool for both computational scientists and clinical researchers to easily generate a graphical results and automatic reports containing the following analyses:
\begin{itemize}
\item overview and visualization of clinical data and biological markers
\item univariate and multivariate classification analysis (logistic regression)
\item univariate and multivariate survival analysis (Cox regression, Kaplan-Meier analysis)
\item correlation analysis
\item statistical tests
\item visualization of markers (boxplots, barplots, volcano plots, forest plots …).
\end{itemize}

This package provides multiple functions to perform machine learning analysis using the `tidymodels` framework. Tasks include: feature selection, plot feature importances, train, cross-validate, apply supervised machine learning algorithms (classification or survival analyses) and unsupervised machine learning, evaluate metrics of predictive performances, compute learning curves.

Initial development was part of the `stats_pioneer` package (also called `pioneerPackage`) and `ml.tidy` evolved as a standalone package only in February 2023.

This R package implements a framework to work with Non-linear Mixed effects models in the context of clinical oncology to predict relapse and survival using longitudinal data.

compOC is a Python implementation of the optimism correction bootstrapping framework for evaluating full machine learning pipelines with or without feature selection. There are 40+ FS methods available, such as Lasso, Stability Selection, Stabl, t-test filtering, hierarchical clustering, recursive feature elimination and others.

This software was built to analyse the PIONeeR (Precision Immuno-Oncology for advanced Non-small cell lung cancer patients with PD-(L) 1 ICI Resistance) data. PIONeeR is a prospective, multicenter study with primary objective being to validate the existence of a hypothetical immune profile explaining resistance to immunotherapy in non-small cell lung cancer patients.

It initially integrated preprocessing, exploratory data analysis, visualization, statistical analysis, feature selection, machine learning and results generation and reporting. Since, exploratory data analysis, visualization and statistical analysis have been promoted to the COMPO-level `compoEDA` package, feature selection and machine learning to the COMPO-level `ml.tidy` package and .

This software corresponds to the very first step of the data analysis, which is the preprocessing, and the very last: generation of results. Some of its functions aim at:
\begin{itemize}
\item preprocessing the data (creation of clinical variables, dictionary, outcome variables, data monitoring and corrections, treatment of the variables types)
\item generating the tools to load the data and metadata
\item computing statistical tests, logistic or Cox regression, or performing a correlation analysis
\item visualising the data (boxplots, barplots, survival curves, ROC curves, volcano plots)
\item providing detailed and interactive statistical reports on the data
\item displaying statistical results and visualisation to interactive dashboards
\item performing supervised and unsupervised machine learning modelling
\item providing detailed and interactive machine learning reports
\item displaying all reports on two websites (one private, one public)
\item automating the production of reports and websites using Gitlab CI/CD
\end{itemize}

This R package is the implementation of a general framework to build and use models of the metastatic process based on the initial model of Iwata et al. (2000). The family of model that can be built describe the metastatic disease with a partial differential equation (pde) on the size structured distribution of the tumors. These models have three components, a function that characterize the growth of the primary tumor, a function that characterize the growth of the metastases, and a dissemination function that decribes how new metastases are produced.

A collection of models implementation for the metamats R package

– Model and simulate (pre-surgical) primary tumor growth and pre- and post-surgical metastatic development. To do so, you define
– A primary tumor growth model (e.g., Gompertz), possibly with a kinetics-pharmacodynamic model for neo-adjuvant (pre-surgery) treatment.
– A dissemination law, that specifies the rate of birth of new metastatic lesions.
– A growth model for metastatic lesions that can be the same of different to the primary growth one.
– Fit the model to data, using a population approach (mixed-effects models).
– Perform simulations to assess the impact of different treatment schedules on metastatic development.
– Assess the predictive value of biomarkers collected, e.g., at surgery, on the mathematical model parameters, using machine learning regression algorithms.

This software fits models of metastatic development to longitudinal data of metastatic sizes and provides simulation and visualization tools for metastatic modeling.

Set of functions for calibration of models of metastatic burden and primary tumor growth to empirical population data

Metamats_core simulates a partial differential equation (PDE)-based model for the time development of a population of secondary tumors (metastases).

SChISModeling aims to analyze SChISM data (Size CfDNA Immunotherapies Signature Monitoring). SChISM is a clinical study that introduces an innovative approach to quantify circulating free DNA in cancer patients treated with immunotherapy. The study's objective is to early predict response to immunotherapy in patients at an advanced/metastatic stage according to these quantitative cfDNA data.

This software corresponds to the very first step of the data analysis, which is the statistical analysis. Some of its functions aim at:
\begin{itemize}
\item preprocessing the data (creation of clinical variables, dictionary, outcome variables, clinical biomarkers, treatment of the variables types)
\item computing statistical tests, logistic or Cox regression, performing a correlation analysis
\item visualizing the data (boxplots, barplots, survival curves, ROC curves, volcano plots)
\item providing detailed and interactive statistical reports on the data
\item simulation for ODE models for mechanistic modeling
\end{itemize}

The app allows the comparison of the gene expression level vs the PAMG which is a transcriptomic signature that describes PDAC heterogeneity as a continuous gradient from pure basal-like (low PAMG) to pure classical phenotypes (high PAMG).

qSingCTools is a web application which allows the pre-processing, analysis and visualization of qPCR Single Cell data. qSingCTools takes a Gene X Cell table of CT values generated by qPCR experiments. Gene expression values were then calculated by applying the y=40-CT formulate. The count values equal to 999 (or missing values) were substituted by values generated from a Normal distribution centered on zero with a standard deviation obtained from the dataset.