Girija Goyal goyalgirijagmail com James Akin jjfourgmail com

Girija Goyal goyalgirija@gmail. com James Akin jjfour@gmail. com

What is Refigure? Reproducibility for every experiment, every figure Outlet and Recognition to young scientists as soon as they complete a replicate or study the literature such that reproducibility information is publically available in real time Crowdsources open access figures from across the web and highlight reproducibility metrics

The reproducibility crisis Open data available from Nature survey of 1576 respondents Penny, Dan (2016): Nature Reproducibility survey. figshare. https: //dx. doi. org/10. 6084/m 9. figshare. 3394951. v 1 Re trieved: 01 18, Oct 20, 2016 (GMT) Modified from Freedman et al (2015) The Economics of Reproducibility in Preclinical Research. PLo. S Biol 13(6): e 1002165. doi: 10. 1371/journal. pbio. 1002165

Data repositories offer a partial solution • Not peer-reviewed • Increased “noise”: millions of pieces of data are submitted but searching and understanding them without annotation is difficult. • There is no way for the published data to be connected to the reproduced literature

t por Cas e re por t Meet Christa, she researchers Zika virus! 2015, Zika virus is still tested in blood Sexual transmission is hypothesized A postdoc at Harvard is studying this question: she knows the literature and is trying to reproduce some experiments

She has two assets: • Her in-depth knowledge of experimental conditions and results • Her reproduced data Neither has an outlet in traditional publishing!

Most websites provide lists of publications that are curated and updated in a conventional way. This may not be responsive enough in emergent crises Science. Open Research. Gate

Re. Figure data dashboard

Who is going to collate these figures? Why? From where? How? • Persona-who and why, target pool, commitment to improving trainee recognition • Where: currently open access, post adoption even journals behind paywalls • Focused on reproducibility but such a meta collection of figures can be curated based on any user defined link • Extension, seamless short interaction • Other use cases • Metrics • Linked to publisher’s website

Formerly mini. Reproducibility Project Open Research Accelerator 2015 Elife innovation Initiative Figure Canvas- an open project launched Mozfest 2016

The majority of our survey respondents are early career scientists Career Level Respondents Percentage Undergraduate 0 0% Grad Student 29 42% Postdoc 32 46% Investigator 8 12% Total 69 100%

These primarily bench scientists (students/postdocs) are in a productive phase of generating results When was the last time you published a paper?

However, an overwhelming amount of their data goes unpublished Q: How much of the data you have generated were NOT used in a final publication? Q: Approximately how much of the data you generated were used to confirm previous findings?

Respondents provided numerous comments about features they would find desirable; they wanted responsive feedback, linking of replicates, and recognition “ autolinks to peer-reviewed pubs if the data is ultimately integrated into such a publication twitter account, links to key words or science themed hashtags Create a "paper" for me based on my submissions Some sort of tracking and/or rating metric(s) to validate how the submitted data is being accessed and received by the scientific community Ensure that if papers replicate experimental designs posted on the database, the original experimenter gets credit. Such a system would be akin to patenting and intellectual property. Peer review - post-publication peer review will help in wider acceptance of the data as well as establishment of the website as worth spending time on, among so many peerreviewed journals that a scientist has to keep tabs on. What do I get: the platform should offer automated feedback to me on my experiment? There must be a reviewer for initial screening of all the data