The  variety of bioinformatics pipelines reported across the 111 papers employed 108 unique pipelines, i.e. sets of bioinformatics tasks carried out in a specific sequence. Three pipelines were used twice; in two of these cases, a group of authors replicated their pipeline exactly, in the other case the pipeline as reported consisted solely of a single step of searching raw reads against a reference set. Although some of these pipelines were similar, with minor modifications to the order, or the addition/removal of a few tasks, the heterogeneity of pipelines is remarkable. There was also high heterogeneity in the number of tasks implemented within each pipeline, ranging between 1 and 18 tasks, with half of the articles reporting fewer than 9 distinct bioinformatic tasks (Fig. 2a). There was no  particular trend in the number of tasks implemented over time (Fig. 2b). The order in which  these tasks were implemented also differed greatly  (Fig. 2c), although there was a tendency for certain tasks to be performed within similar general stages within pipelines, that is, read preparation-based tasks tend to be implemented at the initial steps of the pipelines, followed by filtering-based tasks and data generation tasks (Fig. 3).