Nucleotide conversion sequencing experiments have been
developed to add a temporal dimension to RNA-seq and single-cell RNA seq. Such
experiments require specialized tools for primary processing such as GRAND-SLAM,
and specialized tools for downstream analyses. grandR provides a comprehensive
toolbox for quality control, kinetic modeling, differential gene expression analysis
and visualization of such data.

Documentation page:

Github page:


Digital expression measurements (e.g. RNA-seq) are often used to determine the change of quantities upon some treatment or stimulus. The resulting value of interest is the fold change (often logarithmized).This effect size of the change is often treated as a value that can be computed as lfc(A,B) = log2 A/B. However, due to the probabilistic nature of the experiments, the effect size rather is a random variable that must be estimated. This fact becomes obvious when considering that A or B can be 0, even if the true abundance is non-zero.

We have shown that this can be modelled in a Bayesian framework. The intuitively computed effect size is the maximum likelihood estimator of a binomial model, where the effect size is not represented as fold change, but as proportion (of note, the log fold change simply is the logit transformed proportion). The Bayesian prior corresponds to pseudocounts frequently used to prevent infinite fold changes by A or B being zero. Furthermore, the Bayesian framework offers more advanced estimators (e.g. interval estimators or the posterior mean, which is the optimal estimator in terms of squared errors).

Github page:


Most of our work is based on our Java framework Gedi, a software platform for handling genomic data such as sequencing reads, sequences, per-base numeric values or annotations.

Its main feature are:

  • Comprehensive general purpose software library (Managing iterators, Serialization, Random access I/O, Parallelization, Extension system, JS based template engine, JSON, Strings, Arrays, …)
  • Specialized algorithms and data structures for Bioinformatics (Clustering, maximum scoring subsequences, sequence alignment, string searching, suffix trees, tries, union find, range minimum queries, …) and Statistics (Descriptive statistics, inference, kernel methods, regression)
  • Even more specialized algorithms and data structures for genomic data (Random accessed fasta files, Memory based interval trees, disk based interval trees, space efficient handling of aligned reads, Annotation management, Id mapping,…)
  • Graphical user interface (Genome browser, …)

Github page:


Globally refined analysis of newly transcribed RNA and decay rates using SLAM-seq (GRAND-SLAM) is a computational approach to infer the proportion and the corresponding posterior distribution of new and old RNA for each gene from SLAM-seq experiments.

Project wiki:

Software download: here


Proteogenomic Identification using Stratified Mixture models (PRISM) combines de novo peptide sequencing, efficient string search techniques and mixture modeling based FDR estimation to enable a comprehensive identification of MHC-I immunopeptidomes based on mass spectrometry.

Software download: here


PRICE (Probabilistic inference of codon activities by an EM algorithm) is a method to identify ORFs using Ribo-seq experiments embedded in a pipeline for data analysis

Project wiki:

Software download:


iTiSS (integrated Transcriptional start site caller) is a method to identify transcriptional start sites (TiSS) from various TiSS-profiling experiments with an additional integrative module to combine and remove artefactual TiSS called in single data sets.

Project Readme:

Software download:

HSV-1 Viewer

Since the genome of herpes simplex virus 1 (HSV-1) was first sequenced more than 30 years ago, its predicted 80 genes have been intensively studied. Here, we unravel the complete viral transcriptome and translatome during lytic infection with base-pair resolution by computational integration of multi-omics data. We identified a total of 201 viral transcripts and 284 open reading frames (ORFs) including all known and 46 novel large ORFs. Multiple transcript isoforms expressed from individual gene loci explain translation of the vast majority of novel viral ORFs as well as N-terminal extensions (NTEs) and truncations thereof. We show that key viral regulators and structural proteins possess NTEs, which initiate from non-canonical start codons and govern subcellular protein localization and packaging. We validated a novel non-canonical large spliced ORF in the ICP0 locus and identified a 93 aa ORF overlapping ICP34.5 that is thus also deleted in the FDA-approved oncolytic virus Imlygic. Finally, we extend the current nomenclature to include all novel viral gene products.

To make the annotation and all the obtained data readily accessible to the research community, we here provide our HSV-1 genome browser software. Thereby, viral gene expression and all data can be visually examined from whole genome to single-nucleotide resolution.

Download at zenodo


FERN (Framework for Evaluation of Reaction Networks) is an extensible and comprehensive framework for efficient simulations and analysis of chemical reaction networks written in Java. It includes state of the art algorithms for stochastic simulation and a powerful visualization system based on gnuplot and Cytoscape.

Although FERN consists of more than 100 classes and interfaces, most classes are basically just implementations of one of three major interfaces and abstract classes.

  • The interface Network provides the network structure of the model.
  • The abstract class Simulator performs simulations on a Network. It additionally calls the registered observers during the simulation run.
  • The abstract class Observer traces the simulation progress and creates the simulation output.

FERN is freely available under the GNU Lesser General Public License (LGPL).

Software download: here