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