Kamil S. Jaron

Kamil S. Jaron

genome naturalist;

Data on request

2 minute read

Published:

At the beginning of December 2017 we have decided to do a meta-study/review of several published genomes. I optimistically started to collect data for some supplementary analyses. But then I found that about one fifth of genomes were were available “on request” and about one third of genomes were missing sequencing reads. It was a shock for me, I thought that every paper needs to deposit data to DDBJ / EMBL-EBI / NCBI. This was also the case for genomes published in journals that actually claim that data need to be accessible as PloS Biology.

It was too soon to get depressed. I have decided just to send emails to all the authors asking if they would kindly share data with me. And they actually did, it was a lot of data pulling from file sharing services like dropbox. Some authors sooner, some authors latter, but during January I finally collected all the data I wanted. It felt great, but it was too soon to celebrate.

I am computing on Vital-it, our University cluster. Vital-it data policy says that everything older than one month will be erased1. They send no notifications or warnings, just silently delete your data. Yes, I guess you now see where it goes…

I made a cron job that scans my files and send me notification if anything is going to be erased2. But at beginning of this year we had an big software update including upgrade of operating system, during which they (again silently) erased all cron jobs running on our front-end. I figured out today, because I found that I am missing some reads, but it was too late. Sadly I was not missing just some reads, but virtually all my data. The only files that have survived are those I collected / created this year, which also means that majority of “genomes available on request” are gone as well.

Therefore I would like to ask. What should I do now? Re-think my data management, that’s for sure. Send another round of emails to re-collect all the data? Or accept my defeat and abandon the idea of publishing analyses based on public data? Is there a way how to convince people to deposit data to public archives?

  1. They provide archive service as well, however this data are stored on tapes for eternity and the service is paid per year. ↩

  2. It’s forbidden to write scripts that periodically update timestamps of your files, but monitoring timestamps and sending notifications is not against any rule. ↩

Leave a Comment

You May Also Enjoy

PCI made Google doc temlpates

3 minute read

Published:

Have you heard about PeerCommunityIn, or PCI for short? It’s an organisation, that runs a peer reviewing platform and a diamond-open access journal. The way it works is, you submit a preprit to any of the preprint servers (such as bioRxiv) and then pass the details to appropriate PCI (there are 15 are moment for all sorts of disciplines). There will be a recommender (moreless like an editor), who will manage your preprint, evaluate if it is a credible piece and then send invitations to reviewers. Once you get through the review process, your priperint will get recommended. Once that happens, it’s up to you what you do next, you can either submit the already reviewed preprint to one of the classical journals, or you can right away publish it in the PCI Journal. The initiative is great for reducing work load for reviewers, bringing transparency to the review process and finally brings a sustainable solution to scientific publishing - you should totally check it out.

Testing FastK on a difficult genome

11 minute read

Published:

Once per while there is a new way how to do things and speed up computations in the world of kmer genomics. Sometimes the tricks are simply more efficient algorithms, but sometimes the tricks are shortcuts that don’t do excatly the same thing. Here I would like to dig a bit in the relatively new k-mer counter FastK and compare it with my personal favourite KMC. If you are wondering if it is worth learning new tool, this blogpost might be able to help you make your mind.

How exciting is parthenogenesis in condors?

6 minute read

Published:

Few days ago, the world learned about two California condor chicks that were products of parthenogenesis, a reproduction where unfertilised eggs hatch and develop in an adult from maternal genome only. This IS exciting for multiple reasons, but perhaps not as exiting for conservation of condors and I will try to explain why… It will take a bit of background on parthenogenesis and sex determination, but bear with me.

Morality of academic publishing

10 minute read

Published:

When I was young naive aspiring scientist I did not comprehend all the aspects of publishing. To be honest, I did not think about it much, but for me it’s the same as the climate change, it’s harder and harder to turn a blind eye. The last drop for me was the announcement of Nature’s Open Access option, it’s shocking €9,500, (or $11,390 / £8,290)! How did we come to this? Are we really going to let a private company to drain the already poor scholarly funds by these obscure amounts? And the problem is not just Nature…