Read on Twitter
Here's a twitter version of my @OBF_BOSC #BOSC2020 #BCC2020 talk about @nf_core 
@nf_core is a community effort to collect best-practice analysis pipelines built using @nextflowio. They run on almost any computational infrastructure and bundle all software requirements.
[1/16]
@nf_core is a community effort to collect best-practice analysis pipelines built using @nextflowio. They run on almost any computational infrastructure and bundle all software requirements.
[1/16]
There are @nf_core community guidelines to ensure that all pipelines are consistent and reproducible. To help pipeline users and developers there is a command-line helper tool ( https://github.com/nf-core/tools ) and of course we create and maintain a set of pipelines 
[2/16]
![There are @nf_core community guidelines to ensure that all pipelines are consistent and reproducible. To help pipeline users and developers there is a command-line helper tool ( https://github.com/nf-core/tools ) and of course we create and maintain a set of pipelines [2/16]](https://pbs.twimg.com/media/EdgVpWvWoAA9Oe7.png "There are @nf_core community guidelines to ensure that all pipelines are consistent and reproducible. To help pipeline users and developers there is a command-line helper tool ( https://github.com/nf-core/tools ) and of course we create and maintain a set of pipelines [2/16]")
[2/16]
There are currently 25 pipelines with stable releases. Each release gets a @ZENODO_ORG DOI and tagged @docker
images for ultimate reproducibility.
We aim for one @nf_core pipeline per analysis type, collaborating on adding new features for the whole community.
[3/16]
images for ultimate reproducibility.
We aim for one @nf_core pipeline per analysis type, collaborating on adding new features for the whole community.
[3/16]
Bigger @nextflowio pipelines can have a lot of parameter arguments. @nf_core pipelines will soon all come with a new file describing each parameter in a structured manner 
[4/16]
Pipeline parameter descriptions are written using the standardised @jsonschema format and allow us to automate loads of stuff: validation of inputs when running a pipeline; command line help; even the website documentation - all from a single file 
[5/16]
[5/16]
Large @jsonschema files are difficult to write by hand, so we made a helper tool 
This scans the pipeline code and builds a schema using the parameters it found. It also checks that the schema doesn't have any unrecognised parameters.
[6/16]
This scans the pipeline code and builds a schema using the parameters it found. It also checks that the schema doesn't have any unrecognised parameters.[6/16]
The tool launches a GUI on the @nf_core website. You can group parameters, reorder with drag and drop and write help text using an intuitive interface. You can even choose an icon for each parameter! 
When you're done, the command-line tool saves your updated schema.
[7/16]
When you're done, the command-line tool saves your updated schema.
[7/16]
One of the most exciting things this new @jsonschema allows us to do is to build user interfaces for launching pipelines
Some pipelines now show a new "Launch" button on the @nf_core website, eg. https://nf-co.re/atacseq
[8/16]
Some pipelines now show a new "Launch" button on the @nf_core website, eg. https://nf-co.re/atacseq
[8/16]
Clicking "Launch" takes you to an auto-generated form where you can customise the pipeline parameters for your run. It shows help text and validates your inputs in real-time. 
[9/16]
[9/16]
When you're done, the interface gives you a command that you can copy and paste in your terminal to kick off the pipeline run.
You can also use the command-line tool to run this interface with *any* @nextflowio pipeline (if no schema is found it auto-generates one).
[10/16]
You can also use the command-line tool to run this interface with *any* @nextflowio pipeline (if no schema is found it auto-generates one).
[10/16]
If you're running offline you can do all of this interactively on the command line instead! 
An interactive wizard takes you through the parameters and even renders the #markdown help text in the terminal using #Rich ( https://github.com/willmcgugan/rich)
[11/16]
An interactive wizard takes you through the parameters and even renders the #markdown help text in the terminal using #Rich ( https://github.com/willmcgugan/rich)
[11/16]
Next up! We have a pipeline code linter tool to check @nf_core guidelines. It runs using @github actions, but results can get lost in the logs
Soon,
and 
will be automatically posted as comments on PRs, with descriptions and links to the documentation.
[12/16]
We have a pipeline code linter tool to check @nf_core guidelines. It runs using @github actions, but results can get lost in the logsSoon,
and will be automatically posted as comments on PRs, with descriptions and links to the documentation.[12/16]
We do test pipeline runs using @github actions too, but with tiny datasets due to limited resources.
We wanted to run more realistic full-size datasets, but we were going to need something bigger..
[13/16]
We wanted to run more realistic full-size datasets, but we were going to need something bigger..
[13/16]
Thanks to a grant from @awscloud we are now running automated analysis with full-size datasets for every pipeline release. These use #awsbatch and we can monitor the runs using @SeqeraLabs @nextflowio Tower.
[14/16]
[14/16]
This year @nf_core was published in @NatureBiotech 
https://www.nature.com/articles/s41587-020-0439-x (full text: https://rdcu.be/b1GjZ )
You can see live versions of the plots from the #supplementaries on the @nf_core website: https://nf-co.re/stats
[15/16]
https://www.nature.com/articles/s41587-020-0439-x (full text: https://rdcu.be/b1GjZ )You can see live versions of the plots from the #supplementaries on the @nf_core website: https://nf-co.re/stats
[15/16]
The @nf_core community continues to go from strength to strength - we have hundreds of contributors and a large and friendly @SlackHQ organisation, with dedicated channels for each pipeline.
Come and join us! https://nf-co.re/join
[16/16]
Come and join us! https://nf-co.re/join
[16/16]
If you prefer, you can watch the recorded version of this talk on YouTube:
Feel free to post any questions on this twitter thread, or come and join us in #Slack - https://nf-co.re/join
Feel free to post any questions on this twitter thread, or come and join us in #Slack - https://nf-co.re/join
