Kudos to everyone involved in the project and thanks to Intel's Pandemic Response Technology Initiative for partly funding this project.

We got our #CoVigator pipeline and dashboard for monitoring SARS-CoV-2 published. The dreadful COVID-19 has brought the largest sampling of a viral pandemic with molecular resolution. This comes with challenges and opportunities to understand and predict viral evolution. In the spirit of #opendata we make available a large dataset of clonal and intrahost mutations together with geographical and temporal metadata derived from public resources.

https://www.mdpi.com/2347100 #mdpiviruses @VirusesMDPI

We got our #CoVigator pipeline and dashboard for monitoring SARS-CoV-2 published. The dreadful COVID-19 has brought the largest sampling of a viral pandemic with molecular resolution. This comes with challenges and opportunities to understand and predict viral evolution. In the spirit of #opendata we make available a large dataset of clonal and intrahost mutations together with geographical and temporal metadata derived from public resources.

https://www.mdpi.com/2347100 #mdpiviruses @VirusesMDPI

@elduvelle
I post this like 1000 times a day but for reference for anyone stumbling by: https://jon-e.net/infrastructure/#the-costs-of-infrastructure-deficits

#Python users: beware that the use of PyPI as an attack vector continues and the perpetrators are trying new methods. Targeted packages include Pandas, Beautiful Soup, PyInstaller, Scrapy, TensorFlow, PyTorch, and others.
https://arstechnica.com/information-technology/2023/02/451-malicious-packages-available-in-pypi-contained-crypto-stealing-malware/?utm_brand=arstechnica

Why my news section in the Android client is full of Washington Post articles behind a pay wall? Is this list built from the people I follow?

#mastodonhelp

I've been watching this grow behind the scenes at @SeqeraLabs and couldn't believe it when I first saw the results 🤩 Fusion was built specifically for @nextflowio so is able to do some nifty tricks for performance 🚀 Read how see the full benchmark: https://seqera.io/blog/breakthrough-performance-and-cost-efficiency-with-the-new-fusion-file-system/

#github I have ~80 repos. I'd like to add github `actions` to almost all of these.

But I can't find a page that lists the status of all my `actions` (do they exist? are they passing?) Is there a secret page for this?

#githubactions

Apparently my job here on #Mastodon is to boost a LOT of posts. And I'm OK with that, really. Most of you seem to like it. 🤷‍♀️

But as a periodic reminder for those who might be overwhelmed by it all, you don't have to unfollow or mute me to turn off the firehose.

Simply click on my avatar to see my profile and then click on the "..." menu to choose "Hide boosts from donmelton."

That's it. That will turn off the spigot but still show posts like this one.

Thanks for following! ❤️

Something I learnt about only recently (it may have been on here): for many arxiv preprints, if you replace the "x" in the url with a "5", you will get a nice accessible html :html5: version!

example:
https://ar5iv.org/abs/2203.08489

Works quite well for reading on your phone in the train for example 👍

Data science/ML friends: a question about development

I find that the test data on our dev and test environments is so different from the real user data I work with on prod that it’s difficult to write and test even simple queries/analytic code. Not enough data, too few realistic edge cases.

How do you deal with this issue in your own work?

One of the things that keep breaking my #python packages in #pypi or a #conda repositories like #condaforge or #bioconda is the fuzziness about dependency management.

Let's say my package depends on X, but I don't specify a particular version. All working today, but in a couple of weeks there is a X release which is not compatible.

Even if you set your dependency versions carefully, you cannot control the dependencies of your dependencies...

How do you define your dependencies? Any advice?

One of the things that keep breaking my #python packages in #pypi or a #conda repositories like #condaforge or #bioconda is the fuzziness about dependency management.

Let's say my package depends on X, but I don't specify a particular version. All working today, but in a couple of weeks there is a X release which is not compatible.

Even if you set your dependency versions carefully, you cannot control the dependencies of your dependencies...

How do you define your dependencies? Any advice?

#COVID19 Study:
🔸«The result indicated that the XBB.1.5 had the highest binding affinity level of the spike protein with ACE2 and the longest evolutionary distance of the S gene.»

🔸«XBB.1.5 may be infected farther and faster than can infections of preexisting variants.»

🟧◼️◾️▪️
#SARSCoV2 Omicron XBB.1.5 may be a cautionary variant by in silico study | bioRxiv
https://www.biorxiv.org/content/10.1101/2023.01.18.524660v1

@feargal around three years ago I evaluated both and chose nextflow. I think the main reason was that I compared a badly written snakemake workflow with properly written nextflow workflows... I've recently seen snakemake workflows that definitely look good and simple.

#python peeps: any suggestion for an open source project that is friendly to new contributors?

I’m looking to get involved in the community, I have a fair amount of experience using Python for data science, but am not a software engineer (yet!).

Interested in tools for data science, and data quality/cleaning/governance.

The Pizzigati Prize

Celebrate software developers who create #opensource apps and tools that nonprofit and advocacy groups can put to good use

https://www.nten.org/change/pizzigati-prize

Just published the #CoVigator pipeline in the great #WorkflowHub https://workflowhub.eu/workflows/417 ... any feedback is appreciated as always

Interesting comparison between 2020, 2021 and 2022

#SarsCoV2 #CoVigator

An "artistic" view of #CoVigator #SarsCoV2 dataset showing top 100 lineages and mutations...