EMBL EBI: Bioinformatics for Principal Investigators

On the 17th of June, 4TU.ResearchData’s data train­er, Eiri­ni Zorm­pa, and com­mu­ni­ty man­ag­er, Con­nie Clare, deliv­ered a 3‑hour work­shop on ‘Plan­ning your data man­age­ment’ dur­ing the Euro­pean Bioin­for­mat­ics Insti­tute (EMBL-EBI) course on ‘Bioin­for­mat­ics for Prin­ci­pal Inves­ti­ga­tors’. 

Nine­teen inter­na­tion­al par­tic­i­pants joined the work­shop to learn about data man­age­ment plans (DMPs); why they are need­ed; what infor­ma­tion needs to be includ­ed; and, how to use the DMPon­line tool to cre­ate a DMP. 

You can down­load and reuse our work­shop slides.

A PI’s perspective

The work­shop kicked off with a series of Men­ti ques­tions to get to know the audi­ence. 

Around >76% of par­tic­i­pants admit­ted that their research data man­age­ment work­flow could be improved and/or that they need help, and more than half (56%) of the par­tic­i­pants report­ed that they had nev­er com­plet­ed or heard of a DMP before the ses­sion.  

Ques­tions about repro­ducible research revealed that more than 90% of par­tic­i­pants believed that there is a ‘repro­ducibil­i­ty cri­sis’.

These results close­ly align with those report­ed in the Nature paper, ‘1,500 sci­en­tists lift the lid on repro­ducibil­i­ty’.

The vast major­i­ty of par­tic­i­pants had pre­vi­ous­ly failed to repro­duce some­one else’s work (92%) and/or their own work (86%). They believed rea­sons for this lack of repro­ducibil­i­ty could be due to:

• Poor study design
• Poor data analy­sis 
• Lack of doc­u­men­ta­tion and method­olog­i­cal detail
• Bio­log­i­cal noise and vari­a­tion
• A prob­lem with reagents (wet-lab)
• Incor­rect inter­pre­ta­tion of results
• Study/reporting bias

In gen­er­al, the par­tic­i­pants took a pos­i­tive view on data shar­ing. Around 69% of par­tic­i­pants had used a data repos­i­to­ry to pub­lish their data. They report­ed that pub­lish­ing their data has helped them to: 

• Boost their CV
• Com­ply with pol­i­cy require­ments (e.g. pub­lish­ers) 
• Safe­ly store and pre­serve large datasets 
• Improve long-term find­abil­i­ty of their data
• Share data with col­lab­o­ra­tors 
• Make their data avail­able for reuse

Feed­back gath­ered after the work­shop revealed that 23% of par­tic­i­pants thought that the ses­sion was ‘Good’; 23% thought it was ‘Very Good’; and, 54% thought it was ‘Excel­lent’.

Thank you to the cohort of prin­ci­pal inves­ti­ga­tors who made this work­shop fun and inter­ac­tive!

Spe­cial thanks to Sarah Mor­gan, Jane Reynolds and Patri­cia Car­va­jal from EMBL-EBI for the invi­ta­tion to deliv­er this work­shop, and to Yasemin Türky­il­maz van der Velden (TU Delft) and Mar­ta Teperek (4TU.ResearchData) for the organ­i­sa­tion­al sup­port.

Writ­ten by Con­nie Clare and edit­ed by Eiri­ni Zorm­pa (4TU.ResearchData)