R packages and contributions

Overview

When developing new statistical methods, I usually create R packages that make it easier to apply them. When this is not possible due to time or funding constraints, I upload the R code I used on my github page.

My involvement in the R community so far can be summarized as follows:

• I am the author and mantainer of the R packages neat, ptmixed and pencal;

• I contributed to the development of the R package success (developed and mantained by Daniel Gomon);

• I was one of the organizers of the European R Users Meeting 2020 (e-Rum2020) conference;

• I made minor contributions to the R packages EnrichmentBrowser and markovchain.

Below you can find more details about some of these contributions.

R package neat

NEAT is an acronym that stands for Network Enrichment Analysis Test, a test that can be used to assess relations between sets of nodes (typically genes) in a network.

The method is described in Signorelli, Vinciotti and Wit (2016). NEAT: an efficient network enrichment analysis test. BMC Bioinformatics, 17:352, and it is implemented in the R package neat, available from CRAN.

If you want to learn more about NEAT, you can:

• read the paper that describes the method;

• read the vignette: An introduction to the R package neat;

• visit the package page on CRAN, and check out the package manual.

News: since November 2020, the NEAT test can be computed also using the Bioconductor package EnrichmentBrowser.

R package pencal

pencal is an abbreviation that stands for Penalized Regression Calibration (PRC), a method that can be used to predict a survival outcome given a set of predictors that are high-dimensional and longitudinally measured. PRC is implemented in the R package pencal, available from CRAN.

PRC was first described in Signorelli et al. (2021). Penalized regression calibration: a method for the prediction of survival outcomes using complex longitudinal and high-dimensional data. Statistics in Medicine, 40 (27), 6178-6196. A detailed description of how to use PRC to do dynamic prediction of survival can be found in Signorelli (in review). pencal: an R Package for the Dynamic Prediction of Survival with Many Longitudinal Predictors.

To find out more about PRC and pencal, check out the following resources:

• read the Statistics in Medicine article where PRC was proposed;

• read the article pencal: an R Package for the Dynamic Prediction of Survival with Many Longitudinal Predictors;

• visit the package page on CRAN, and check out the package manual.

R package ptmixed

ptmixed is an abbreviation that stands for Poisson-Tweedie generalized linear mixed model, a mixed-effects model that has been developed to flexibly model longitudinal count data that feature overdispersion, zero inflation and/or heavy tails.

The model is described in Signorelli, M., Spitali, P., Tsonaka, R. (2021). Poisson-Tweedie mixed-effects model: a flexible approach for the analysis of longitudinal RNA-seq data. Statistical Modelling, 21 (6), 520-545, and it is implemented in the R package ptmixed, available from CRAN.

To find out more about ptmixed:

• read the article that describes the methodology behind ptmixed;

• read the vignette: An introduction to the R package ptmixed;

• have a look at a short 5 minute talk about ptmixed that I presented at eRum2020;

• visit the package page on CRAN, and check out the package manual.

R package success

success (short for: Survival Control Charts Estimation Software) is an R package that implements methods to compute and visualize the following charts for quality control of survival outcomes:

1. funnel plot;
2. Bernoulli CUSUM;
3. Biswas & Kalbfleisch’s CUSUM (BK-CUSUM);
4. Continuous Time Generalized Rapid Response CUSUM (CGR-CUSUM).

Note that the funnel plot and Bernoulli CUSUM require a dychotomization of the survival outcome into a binary outcome, whereas the CGR-CUSUM and BK-CUSUM can properly handle the survival outcome.

To find out more about success:

• visit the package page on CRAN;

• check out this article that illustrates how the package works and what you can do with it;

• visit the package page on CRAN, and check out the package manual.

European R Users Meeting 2020 conference (e-Rum2020)

I was one of the organizers of the European R Users Meeting 2020 conference (e-Rum2020), both as member of the organizing committee and as responsible of the promotion of the conference.

e-Rum2020 was originally planned as a physical event to be held in Milan in May 2020. However, due to the COVID19 pandemic we decided to turn the event into a free virtual conference - the very first fully virtual R conference ever!

If you would like to know more about e-Rum2020, you can:

• read the article e-Rum2020: how we turned a physical conference into a successful virtual event, published in The R Journal, in which we summarized our adventure as conference organizers in the middle of a pandemic outbreak;

• check out the website of e-Rum2020;

• have a look at my curated list of e-Rum2020 resources.