Exploring how technical and social advances can overcome the hurdles of reproducibility in the life sciences

Samota, Evanthia (2022) Exploring how technical and social advances can overcome the hurdles of reproducibility in the life sciences. Doctoral thesis, University of East Anglia.

[thumbnail of 2023Kaimaklioti-SamotaEPhD.pdf]
Download (7MB) | Preview


Reproducibility is an important element of robust science. Reproducibility is vital for the better understanding, validation, and reuse of research. For the past three decades, literature on the reproducibility crisis has increased, highlighting how researchers cannot rerun the analysis of other researchers and reach the same results.

To achieve research reproducibility, there is a need for the raw data, the analysed data (including negative data), the code and complete methodology, the detailed analysis protocol, the standardised data and metadata annotation, to be well documented, and shared so they can be easily accessed by other researchers who wish to reproduce the work. However, with the increasing volume of data, given the advances in life sciences technology, there are still issues with reproducibility, despite the many tools, data and code-sharing mandates created to manage the irreproducibility problem.

This thesis explores how technical and cultural advancements can address irreproducibility issues. The study aims to understand how tools and cultural factors (e.g., training, incentives and rewards for reproducible research practices) promote research reproducibility. A survey of 251 researchers from various backgrounds reported their knowledge of reproducibility issues in the life sciences, their ability and motivation to reproduce research studies and their opinions on the interactive representation of research results (in interactive figures instead of static figures) within journal articles.

Additionally, this thesis investigates how interactive figures could reproduce computational experiments presented in the figures and their benefits and limitations in improving research reproducibility.

Lastly, this thesis presents a software prototype, Deus ex machina, which automatically annotates articles and their metadata with standardised semantic information (plant ontology terms and IDs). Deus ex machina computes a reproducibility metric score that evaluates the reproducibility status of papers, ultimately recognising reproducible research. The software can thereby serve as a means of promoting a more reproducible research culture.

Item Type: Thesis (Doctoral)
Faculty \ School: Faculty of Science > School of Biological Sciences
Depositing User: Chris White
Date Deposited: 19 Jun 2023 10:06
Last Modified: 19 Jun 2023 10:06
URI: https://ueaeprints.uea.ac.uk/id/eprint/92424


Downloads per month over past year

Actions (login required)

View Item View Item