Episode 2 - Reproducibility with Dr Tim Errington

In this episode, we discuss reproducible and open science practices or the lack thereof. Research is both innovative and self-correcting - or maybe not? Are we losing robustness by constantly aiming for novelty and breakthrough results? In 2011, the Center for Open Science set out to estimate the robustness of findings in the field of cancer biology by trying to replicate 50 landmark papers published between 2011-2012. We have a conversation with Dr Timothy Errington, Head of Research at the Center for Open Science, and manager of the Reproducibility Project: Cancer Biology. We learn more about replicability and factors that influence it, challenges encountered during the project, and what researchers can do to change the lack of replicability in scientific publications and their attitude towards errors. // This episode was hosted by Eric Danner @EricDannerBio and produced by Joachim Fuchs @JoachimFuchs_. Sound editing and original music by Giorgio Cattaneo: linktr.ee/exoplanetrec // To keep up with episodes and engage in the discussion follow us on twitter @DoSciDiff // For full show notes, transcripts, resources and more information on this project brought to you through the BIH QUEST Centre for Responsible Research visit: www.bihealth.org/en/quest/projects/spokes https://doingsciencedifferently.start.page // You can find our guest Tim Errington https://www.linkedin.com/in/tim-errington-558a928/ https://scholar.google.com/citations?user=TNiKPdoAAAAJ&hl=en https://www.fetzer-franklin-fund.org/media/tim-errington/ // You can learn more about the Center for Open Science: https://www.cos.io/?hsLang=en

Read specifically about the reproducibility in cancer biology project: https://www.cos.io/rpcb

Read all the project's papers: https://elifesciences.org/collections/9b1e83d1/reproducibility-project-cancer-biology

Read Tim's major paper, mentioned in the podcast, in which 50 experiments from 23 preclinical cancer biology papers were replicated: https://elifesciences.org/articles/71601 // Extra resources: The Bayer/Amgen studies: https://doi.org/10.1038/483531a https://doi.org/10.1038/nrd3439-c1

The power debate - small sample sizes overestimate effects: https://doi.org/10.1038/nrn3475 https://doi.org/10.1038/nmeth.2738

Protocols.io is an open access and free repository for detailed protocols: https://www.protocols.io/

Executable Research Articles (ERAs), articles published in the open access journal eLife 'enriched' with data, interactive figures as well as live code creating the figures: https://elifesciences.org/for-the-press/eb096af1/elife-launches-executable-research-articles-for-publishing-computationally-reproducible-results