owned this note   changed 4 years ago  
Published Linked with GitHub open-life-science/ols-2/cohort-calls/week-10.md

Week 10

tags: OLS-2 coworking call project-leads

Image Not Showing Possible Reasons
  • The image file may be corrupted
  • The server hosting the image is unavailable
  • The image path is incorrect
  • The image format is not supported
Learn More →
Cohort call

Image Not Showing Possible Reasons
  • The image file may be corrupted
  • The server hosting the image is unavailable
  • The image path is incorrect
  • The image format is not supported
Learn More →
Before this meeting

After this meeting:

  • Think about and plan to apply FAIR principles to your project in your own time. Resources for more information have been linked in the notes.

Table of contents

Image Not Showing Possible Reasons
  • The image file may be corrupted
  • The server hosting the image is unavailable
  • The image path is incorrect
  • The image format is not supported
Learn More →
Cohort call

🌍 Cohort call

OLS-2 week 10: Open Science III: Next steps - applying FAIR research principles

Call time: 05 November 2020, 09:00 - 10:30 GMT/UTC (see in your time), Duration: 90 minutes

Hosts: Emmy, Malvika, Yo

Syllabus: OLS-2 Week-10

During this week’s cohort call, we will discuss:

Next steps in Open Science: applying FAIR research principles.

FAIR stands for findable, accessible, interoperable, reusable. The ideal is that research should be “as open as possible, as restricted as necessary.”

  • FAIR Data
  • FAIR Software
  • FAIR Training

Roll call

Breakout room preference:

Please edit your Zoom name (click on the three dots on the top right of your video) and add one of the following letters in front of your name.

  • W for Written Discussion Breakout Room
  • S for Spoken Discussion Breakout Room

This will help us assign you to the breakout room with format of your choice (Read more about why we run different format of breakout rooms):

Name (pronouns - optional) / Project / social handles (twitter - t, GitHub - gh, etc.) / emoji mood

  • Yo Yehudi / OLS / t: yoyehudi g: yochannah 😴
  • Pradeep Eranti (he/him)/ Open platform for Indian Bioinformatics community / t: @pradeeperanti
  • Emma Karoune/ Open Science in Phytolith Research/ t: ekaroune/🤓
  • Emmy Tsang (she,her)/ OLS, TU Delft / t:emmy_ft /
  • Fotis Psomopoulos / Institute of Applied Biosciences (INAB|CERTH) / t: @fopsom , gh: @fpsom / 🙃
  • Arielle Bennett (she/her) / OLS mentor / @biotechchat / 🥴
  • Sam Van Stroud / Turing Data Stories / @samvanstroud
  • Georgia Aitkenhead/The Alan Turing Institute/GeorgiaHCA/
  • Kate Simpson / Open retrofit protocol to evaluate housing decarbonisation / t: Dr_KateSimpson/ g: @KateSimpson
  • Mallory Freeberg / OLS mentor / @malloryfreeberg
  • Jelioth Muthoni/github:@Jelioth
  • Camila Rangel Smith / Turing Data Stories/ t: @camilarangels
  • Jez Cope (he/him) / OLS Mentor / @jezcope
  • David Beavan (he/him) / The Alan Turing Institute / @DavidBeavan
  • Paul Owoicho / Embedding Accessibility in The Turing Way Open Source Guidance / g: paulowoicho, t: ogbonokopaul / 🤓
  • Piv Gopalasingam / he/him / OLS mentor / t: @cascade21 / gh: @PivG 🌁(foggy)
  • Melissa Burke/ she/her/ OLS mentor for Chronic Learning/ t: @burkemlou
  • Ismael Kherroubi Garcia / Guide for Ethical Research / t: @IsmaelKherGar
  • Philippe Rocca-Serra: he/him / Data Readiness Group/Uoxf/t:@Phil_at_OeRC
  • Anelda van der Walt: she/her/ Talarify/ @aneldavdw
  • Laura Carter / Turing Way Guide for Ethical Research / t: @LauraC_rter
  • Karega Pauline/ The-Hub-Portal/gh: @karegapauline
  • Markus Löning / he/him / sktime / t: @mloning_ gh: @mloning
  • Bailey Harrington / Chronic Learning / t: @ChronicLearning, @baileythegreen; gh: @baileythegreen
  • Harriet Natabona/ covid19 open science communication/ g: @natty2012
  • Cooper Smout / Project Free Our Knowledge / @coopsmout @projectfok

❄ Icebreaker question

If you could meet one person from the past (before living memory, say the ~1930s or earlier), who would it be?

Name / answer

  • Yo / Marie curie
  • Ismael / Plato
  • Piv / Charles Darwin
  • Emma / Rosalin franklin
  • Kate / John Ruskin
  • Georgia / Hatshepsut
  • Arielle / Ada Lovelace
  • Sam / Newton
  • David B / Tupaia (Polynesian navigator) +1 (Markus) https://en.wikipedia.org/wiki/Tupaia_%28navigator%29
  • Fotis / Pierre-Simon Laplace (https://en.wikipedia.org/wiki/Laplace_transform)
  • Mallory / Ada Lovelace (Arielle - let’s double-date!) Yes let’s do it!
  • Camila / Florence Nightingale
  • Paul / Abraham Lincoln
  • Melissa / Mary Mallon (Typhoid Mary) because I just read a book about her
  • Laura / Mary Wollstonecroft
  • Philippe / Ignaz Semmelweis

🗣️ Welcome!

Yo: (⏰4 mins)

Call recording reminder:

  • Please note that this call will be recorded
  • The video will be available on the YouTube channel in the next few days
  • Turn on your webcam if you don’t mind being in the video (or off if you do!)

Participation reminder:

🖥 FAIR DATA

Emmy (⏰ 15 mins)

Slides: https://docs.google.com/presentation/d/1ER_ZQ_Fe_PFWBPrP87dU6jRaPT2dlB4hCu-wGXJwnBA/edit?usp=sharing

Presenter: Phillipe Rocca-Serra

  • Contact:
    • Email:
    • Other info:
  • Resource recommendations:

Notes

  • FAIR in the context of Data - data for machines!
  • Data generation and data availability gap - loads of data being generated, but missed opportunities if a large part of this data is not FAIR!
    • The data is around, but we don’t know/can’t see/can’t use it.
  • Having metadata makes things findable and more easily understandable, but it’s also important to check that metadata to make sure that they are correct
  • FAIR in practice - many stakeholders need to be onboard to provide the necessary tooling and infrastructure to make it possible, community standards to provide guidelines, and a cultural change to turn this into a norm

Open Q & A time

  • Yo - I loved the missing labels on food cans comparison! +3
  • Will these slides also be made available after the call?
    • Thank you, Phillipe!
  • I didn’t know you could reserve a DOI “so easily” on zenodo..🤩

🖥 FAIR Training

Emmy (⏰ 15mins)

Slides: https://f1000research.com/slides/9-1294

Presenter: Melissa Burke

Notes

  • Why we need FAIR training materials:
    • It’s really difficult to find training materials - despite that there are loads of people doing training on similar topics, and it’s difficult to understand how they have adapted them
    • Sometimes even when you find them, you can’t access them ):
  • 10 steps for making training FAIR! 👆
    • Share: where to share, what metadata to share
    • Describe properly: bioschemas make it easier for search engines to find and index content and materials; use keywords from ontologies for consistent descriptions
    • Give unique identity - to ensure that contributors can get correct credit!
    • Register online
    • Define access rules
    • Use interoperable format - consider the pros and cons of different formats, some are more adaptable than others
    • Make reusable for trainers - provide detailed metadata; include what, why and how, preferably with structured metadata; and add a license! CC licenses are great for this
    • Make usable for trainees- for folks who want to self-learn
    • Welcome contributions - be welcoming, use a warm tone and have rules for participation and contribution
    • Keep materials up-to-date!

Open Q & A time

  • What types of licenses would you recommend for reuse of materials by trainers? Nevermind, answered! It depends on the purpose of your training materials and any institutional/IP constraints you may have. Creative Commons licences are commonly used for example CC-BY-SA or CC-BY. You can read more about them here https://creativecommons.org/licenses/ Also have a look at https://creativecommons.org/about/cclicenses/
  • Should we be thinking about sharing templates for training materials to help support standardisation /FAIRificiation?
    • There are so many flavors of training - by templating we could be constraining the types of training being provided… we want to maintain that flexibility. The important thing is to let people know the considerations that they should have when developing training materials
    • It is possible to work towards consistent ways of describing training materials e.g. by using metadata standards that indicate what type of information should be provided with training materials and by using consistent keywords to tag materials
  • Thanks for the great presentation. Looking forward to the expanded version of the “10 rules”

✏️ Reflection exercise:

Yo (⏰ 10 minutes) 40

Open science is all very well but how do you make it FAIR in practice?

Prompt: Choose one of the aspects of FAIR (either findable, accessible, interoperable, or reusable) and reflect on what you will have to do in your project to maintain this aspect of your choice. Read other’s comments, +1, add more.

To make my project […] I will have to ensure that:

Findable

  • Kate S: The data is placed within a ‘data catalogue’ (thanks Jez!) for easy finding
  • It is shared in places where potential contributors can find it - maybe Twitter and not just GitHub?
  • Emma K - Writing a data paper - this helped me a lot in my last project as your dataset is very findable but it also allows you to describe it in more detail than you would normally put in a research article so therefore includes metadata too.2
  • Dataset is published openly somewhere that the people who are likely to find it useful are also likely to be familiar with (e.g. GitHub)
  • Setting up a Zenodo account and uploading materials so that they have a persistent DOI +3

Accessible

  • Make sure datasets are available in both machine and human readable formats (e.g. cvs and pdf) + 3
  • Good documentation around data + 1
  • Give appropriate context to training materials such as who it was designed for (audience information such as career level, subject area, beginner/intermediate/advanced training, date of original training, why the training took place)
  • Use headings in my docs +1
  • If using GitHub, turn your repository into a GitHub Pages site (there are templates you can use without much effort); allows more people to interact with it at a superficial level without needing to be comfortable with GitHub.

Interoperable

  • Make sure source code compiles on different operating systems or distribute pre-compiled files for different operating systems
  • Using standard categories for data (e.g. same rating scales in psychology questionnaires), so that it can be combined with other datasets
  • It is written in a language that makes sense in other fields (ethics, philosophy, science and openness are interoperable, but there are specific concepts for each!) + 1 +1
  • Emma K - Using standard reference for geographic names such as geonames.
  • David B / adopt and use establishes ontologies, failing that, at least express your data in well supported ways e.g. XML / JSON. I’ve seen some database dumps from crappy old software in my time that cannot be joined to anything else
  • Laura C / Translation into other (human!) languages

Reusable

  • Sam / Making sure code is properly commented! Proper commenting takes time but goes a long way when someone in the future wants to reuse or recreate some functionality you have already provided. +1
  • David B / Choose as permissive a licence as possible +3
  • David B / Explicitly encourage contributions and a community around your artefact +1
  • Laura C / Keep the information up to date (and/or clearly mark when it is no longer relevant or useful) - it might even be helpful to have ‘last updated’ information and a target date for updating specific pages +1
  • Use something like conda, Docker, et cetera, to make the exact environment needed for software available to users. (Conda saves details about dependencies, resolves conflicts, and creates a virtual environment that can be shared, but still involves installation on the user’s system. Docker makes something does not require the user to install anything; it is completely self-contained.)
  • Write software so that it is ‘user-proof’: easy to use, hard to mess up. Set sensible defaults; add interactive bits where it alerts the user to a problem and provides an opportunity to fix it; don’t require bizarre input formats and make output formats commonly-used things, too.

Shared reflection, comments, Open Q & A

  • Kate: Loving reading everybody’s ideas! Really helpful, thanks :) +2 +1 +1
  • @Anelda van der Walt: very true -> tools need improving. May be opportunity for funders to support communities both in terms of sw development but also more training in data management.
  • Piv: I don’t have the technical skills (yet) but generally play to my strengths - describing materials (making them human readable!) and supporting/encouraging trainers to FAIRify. Reading all the comments above has been enlightening, I never knew what a data paper was until now! Thank you @Emma K!

🖥 FAIR Software

Emmy (⏰ 15 minutes) 55

Slides

Presenter: Fotis Psomopoulos

Notes

  • Doesn’t make sense to sit around data without software to do anything with them! Research community is increasingly relying on software.
    • Researchers already have a lot to think about, so it’ll be nice to not have to worry about FAIR software
  • Similarities between data and software
    • Both are not cited currently ):
    • Both can be used as the basis for other things to build upon them
  • Differences:
    • Software has a loootttt more dependencies
    • Reuse comes in many flavours - execute/run, use, etc

Open Q & A time

Breakout Discussion

(⏰ 15 minutes) 70

Sharing your notes from the reflection exercise: Now that you know about FAIR principles, and how it can be applied across different research components, share your thoughts from the reflection exercise about how you could apply F, A, I, or R to your work. Would this method apply to others in your group? Why or why not? Further prompt for a written and spoken discussion room is provided below.

Breakout room reminder: if you need assistance in your breakout room, please click the ‘Ask for Help’ button at the bottom of your screen

⏰ Shared breakout room time: https://cuckoo.team/ols-2

👥 Break-out room (12 minutes, ~3-4 rooms):

Breakout room 1, written: ✏️

  • Person 1: Copy and paste your reflection/shared insights from the previous exercise.
  • Others: comment underneath in a sub-bullet-point - would it apply to your work? +1. Provide a written note on why? or why not?
  • Once everyone has had a chance to comment, repeat (1) with the next person in the group.
  • In the last 1.5 minutes, take important notes below and add any question for the complete cohort

Roll call:

  • Person 1: Sam
  • Person 2: Jelioth

Notes

Reflection notes from person 1: Sam

  • In our project, we produce notebooks that teach people how to do data analysis on some open data using Jupyter noteboooks. Uploading the data to Zenodo sounds like a good way to ensure that the data remains accessible and findable in the future. At the moment we cited the data, but perhaps it would not remain available from the original source in the future.
  • “Give appropriate context to training materials such as who it was designed for (audience information such as career level, subject area, beginner/intermediate/advanced training, date of original training, why the training took place)” This is also a really useful idea for our project! It definitely adds accessibility as someone would be able to access material that is most suited to them.
    • Providing access of data to the right people is very useful for it saves on time, just like our project we aim to make immunological data more accessible.
  • We will aim at creating our project accessible so that other people can find it and be of impact in the community

Notes from person 2: Jelioth

  • My reflection is that data implies : it’s not how much good at gathering data if you don’t have the ability to analyze it. In our project we aim to gather data on Immunology, hence providing an easy access of information to scientists. Moreso with the current surge of Covid-19 publications. Thus utilizing a software management plan will be very important.
    • What kinds of things need to go into the software management plan? This sounds like a very useful thing to have! I think in general we are already trying to manage the code we write in our project so that it adheres to certain standards, but perhaps we could do a better job of specifying these clearly :)
    • Sure, it's very useful. Software management is an area that we need to learn more about since it’s not an area of speciality. So that we can be specific and clear about it.
    • It’s good progress that you are already trying to manage the code you write in your project

Breakout room 2, written: ✏️

  1. Person 1: Copy and paste your reflection/shared insights from the previous exercise.
  2. Others: comment underneath in a sub-bullet-point - would it apply to your work? +1. Provide a written note on why? or why not?
  3. Once everyone has had a chance to comment, repeat (1) with the next person in the group.
  4. In the last 1.5 minutes, take important notes below and add any question for the complete cohort

Breakout room 3, Spoken: 🗣

  1. Person 1: Share main points from the previous reflection/shared exercise.
  2. Others: discuss if it would apply to your work? why? or why not?
  3. repeat (1) with the next person in the group.
  4. In the last 1.5 minutes, take important notes below and add any question for the complete cohort

Roll call:

  • Person 1: Emma Karoune
  • Person 2: David Beavan
  • Person 3: Anelda vd Walt

Notes

  • Writing a data paper - this helped me a lot in my last project as your dataset is very findable but it also allows you to describe it in more detail than you would normally put in a research article so therefore includes metadata too
  • David B / when working on private datasets (that are not open), how can we make sure our research, software and derived data can be fully shared?
  • FAIR - doesn’t mean data have to be fully available online always, but at least metadata about the data should be FAIR so that others can be aware and can approach project owners

Breakout room 4, Spoken: 🗣

  1. Person 1: Share main points from the previous reflection/shared exercise.
  2. Others: discuss if it would apply to your work? why? or why not?
  3. repeat (1) with the next person in the group.
  4. In the last 1.5 minutes, take important notes below and add any question for the complete cohort

Roll call:

  • Person 1: Pradeep
  • Person 2: Markus
  • Person 3:

Notes

  • We went a bit off topic but we had a very insightful discussion!

Open Q & A time

  • How to convince “pure” biologists to participate in the FAIRness movement of the datasets (Microarray, RNA-seq, …)?
    • For example, an effort started 20 years ago (doi:10.1038/ng1201-365) and still there is room for convincing. How to create awareness for the rest of the unfollowers?

Breakout room 5, Spoken: 🗣

  1. Person 1: Share main points from the previous reflection/shared exercise.
  2. Others: discuss if it would apply to your work? why? or why not?
  3. repeat (1) with the next person in the group.
  4. In the last 1.5 minutes, take important notes below and add any question for the complete cohort

Roll call:

  • Person 1: Bailey Harrington
  • Person 2: Laura Carter
  • Person 3: Philippe Rocca-Serra

Notes

  • R - Laura C / Keep the information up to date (and/or clearly mark when it is no longer relevant or useful) - it might even be helpful to have ‘last updated’ information and a target date for updating specific pages
  • I - Laura C / Translation into other (human!) languages
  • F - If using GitHub, turn your repository into a GitHub Pages site (there are templates you can use without much effort); allows more people to interact with it at a superficial level without needing to be comfortable with GitHub.
  • R - Use something like conda, Docker, et cetera, to make the exact environment needed for software available to users. (Conda saves details about dependencies, resolves conflicts, and creates a virtual environment that can be shared, but still involves installation on the user’s system. Docker makes something does not require the user to install anything; it is completely self-contained.)
  • R - Write software so that it is ‘user-proof’: easy to use, hard to mess up. Set sensible defaults; add interactive bits where it alerts the user to a problem and provides an opportunity to fix it; don’t require bizarre input formats and make output formats commonly-used things, too.
  • Need put pressure on manufacturers to make it easier to e.g. have equipment output information in a way that is FAIR
  • Funders need to be convinced as well - both to fund the time and resources to e.g. develop software in a FAIR way

Breakout room 6, written: ✏️

  • Person 1: Copy and paste your reflection/shared insights from the previous exercise.
  • Others: comment underneath in a sub-bullet-point - would it apply to your work? +1. Provide a written note on why? or why not?
  • Once everyone has had a chance to comment, repeat (1) with the next person in the group.
  • In the last 1.5 minutes, take important notes below and add any question for the complete cohort

Roll call:

  • Ismael Kherroubi Garcia
  • Paul Owoicho
  • Camila Rangel Smith
  • Kate Simpson

Notes

Reflection notes from person 1: Ismael K.G.

  • Findable: It is shared in places where potential contributors can find it - maybe Twitter and not just GitHub?
    • I’ve been thinking the same thing, we need to find more contributors. But it is not just publishing 1 tweet, is designing a campaign to make sure it reaches the right people, and that i find it more difficult. Do you have a strategy thought out?
      • 100%! But unfortunately not. We need a social media manager, haha. I jest. Something we’ve considered is streaming our work/discussions live through Twitch. One of the team shared this is done by coders, so it might be an idea. But as you say, it needs a strategy!
        • I would love to hear more about it, maybe we have a social media for finding contributors brainstorming session.
  • Kate:I would liek to see this but it is early stage in energy and building research
  • Interoperable: It is written in a language that makes sense in other fields (ethics, philosophy, science and openness are interoperable, but there are specific concepts for each!). For context, I am in a team writing about ethical research in The Turing Way.
    • +1. This would work for my project as well but from an accessibility perspective Our goal is to ‘embed’ accessibility in the Turing Way, and translation is key
    • Kate:There is some work in relation to this is building modelling, but I think it may be a step beyond what I hope to do in the time frameme here.

Reflection notes from person 2: Paul Owoicho

  • Findable: Ensure that the project is publicly available on Github
    • +1 , maybe also giving it a DOI to the repo could help.
  • Accessible: Making it easier for new contributors to get involved by creating good first issues or similar
    • +1 to the need for good first issues, I love that!

Reflection notes from person 3: Camila

  • Reusable: We want our project to provide materials that can function as an interesting insight from data (blog post type) as well as a material that people new to data science/data analysis can use to learn. These two functions need different output formats (jupyter notebooks in binder and fastpages in the form of a publish blog).
    • +1 to the variety of output formats! A blog can make it all a lot more accessible too (I might share this idea with my team!)

🗣️ Closing

(⏰ 5 mins)

  • Think about and plan to apply FAIR principles to your project in your own time. Resources for more information have been linked in the notes.
  • Next calls

Feedback

What worked?

What didn’t work?

  • Zoom was very unstable this week - for me at least!

What would you change?

What surprised you?

Reference: Open leadership Framework, Mozilla Open Leaders 6 & 7, Open Life Science 1 License: CC BY 4.0, Open Life Science (OLS-2), 2020

Select a repo