 Hi, thanks for joining. My name is Phil Jules and I work at the National Genomics Infrastructure, part of Sidelife Lab in Sweden. Today I'm going to talk to you about NFCore, a community effort to collect a curated set of analysis pipelines built using NextFlow. NextFlow is a workflow manager that allows you to change steps of computational analysis together. It orchestrates all of the backend across multiple different compute infrastructures and also handles all of the software packaging for you. NFCore is a community project. We have guidelines for best practices, we have a Python package with helper tools for users and for developers, and we have a core set of pipelines. We try to have one pipeline per data type. Currently we have 25 different release pipelines, all of which have DOI and tagged software releases, so they're highly reproducible. One new feature for NFCore pipelines is just being finished now, it's pipeline schema. NextFlow pipelines have a special variable type called pronouns. When you run a pipeline, it behaves as you expect, and if you specify a double dash with that variable name, it'll be overwritten. These are used extensively to customise the way in which pipelines are run. However, pipelines can have a large number of parameters, making it difficult to know what's available, also making documentation difficult. We're adding a new file to our NFCore pipelines to describe these parameters, give extra metadata such as their type, and also description and help text. This allows us to automate a large number of things around the pipeline, validate user inputs when people run pipelines, but also build health and documentation as well as user interfaces. Building large JSON files is not fun, so we've written a helper tool to auto-generate these from the nextFlow code. It launches a web interface where you can graphically organise and describe the different parameters in your pipeline. This makes it fast and easy to add a lot of metadata to your schema. Because the schema are totally standalone, we can start to use them at other places, such as on the web. Here, I can find a pipeline I'm interested in, click through, and just hit launch. I'm taken to a web page which is built dynamically using the schema, and I can go through all of the different pipeline parameters, fill in what's important, and read help when necessary. All of the inputs are validated, so when I get to the bottom, if I click launch, and something looks wrong against the schema, I'll be immediately told. Once I click launch, my inputs are saved, and I can have a very simple command which I can paste into a terminal. This will pull all of the inputs, build the nextFlow command, and if I'd like to, you can also run it for me. Alternatively, I can just copy all these variables and run nextFlow directly. To ensure that all NFCore pipelines adhere to the guidelines, we make extensive use of code-linting. This checks the pipeline code to make sure it passes all of the different lint tests. This is run automatically whenever you push a change or open a pull request on GitHub. However, these warnings get varied in the logs if none of the tests fail. Feature that's coming soon is an automated comment to the pull request showing what tests have passed and failed with details and links to documentation. Every NFCore pipeline has a minimal test dataset. However, we'd really like to run on full-scale datasets as well. The infrastructure available on GitHub Action's continuous integration is not big enough, so in order to do this, we started a mini project called WS MegaTests. This uses GitHub Actions but launches a full-scale run using AWS Batch. We hope to soon make all of this data available on the NFCore website so you can browse it when exploring pipelines and also use it for benchmarking. One really nice feature is that we can monitor these runs in a new nextFlow tower tool, showing us exactly what's happening with the run and even allowing us to monitor the execution log in real-time. I'm also happy to report that we recently had an NFCore manuscript accepted in nature biotechnology. It's a nice short paper that describes the framework and has large supplementaries describing how we manage the community and the pipelines. The NFCore communities continue to grow very fast over the past few years and you can visit the NFCore website to find lots of stats and graphs about this. But now we're many hundreds of users who are active on this on a daily basis. We're spread all over the world and we'd love for you to join us. We use GitHub for all of our work, we communicate over Slack and we have an active Twitter and YouTube channel. I'd like to thank all of the NFCore contributors and I'd like to thank you for listening and also thanks to the conference organisers.