0_summary.tex

\chapter*{Summary}

The advent of high-throughput sequencing gave rise to exploratory data analysis in biological research. 
As bioinformatics analysis becomes increasingly central to generating and investigating biological hypotheses, analysis workflows have grown more complex.

This thesis aims to provide tools that enable bioinformaticians to gain biological insight at three different levels: a tool for the exploratory analysis of single-cell RNA-seq data, an infrastructure tool facilitating interoperability in the field, and a visualization method offering insight into benchmarking data.

Single-cell RNA sequencing enables the measurements of gene expression in individual cells, allowing researchers to identify structure based on differences in gene expression. One way to approach this is by using trajectory inference, an analysis type that orders cells according to an assigned pseudotime. However, little guidance exists on how to approach analyzing pseudotime across samples. For this reason we developed dynchro, a tool for exploratory comparative trajectory analysis.

Second, we address interoperability between ecosystems and programming languages in single-cell tool development. The field centers around three ecosystems spanning two programming languages. Interoperability between these ecosystems has been unstable, with existing conversion tools offering limited functionality and unreliable performance. We implemented a rigorously tested software package capable or reading in, writing out, and converting a previously Python-specific data format, greatly improving interoperability in the single-cell field.

Finally, we developed funkyheatmap, a widely used visualization tool for benchmarking results, facilitating the interpretation of complex method comparisons.