Project costing for the UKB RAP
Hello I am a PhD student in my first year of study, hoping to use the UKBiobank dataset to complement my research. I will be applying for the UK Biobank Platform Credits which I understand is £1000 per year. Can anyone please advise me as to whether this is likely to cover my cloud costs for the extraction and data analysis for my project? I appreciate that there are many variables that may affect the overall costs, but due to limited funding for my PhD I cannot afford for the costs to spiral. Is is realistic to think I can manage the analysis without it costing me a lot more money than the credits would provide? Any advice would be gratefully received, particularly if anyone can give me an indication of how much their project cost. Thank you.
Comments
2 comments
Hi Christine,
I am not an expert on costs, and these are guesses based mainly on my own use of the UKB-RAP, plus anecdotal information.
If you are using raw images or Whole Genome Sequencing data, you could very easily use more than £1000.
If you are only using tabular tier-1 data, it should be possible to manage well within the credits. I suggest you keep a note of what you've done on UKB-RAP each week, and how your account total has changed, at least for the first few weeks. Please note that if you do something silly such as starting several RStudio sessions and leaving them running indefinitely you could still spend a lot even with tier-1 data.
If you are using Whole Exome Sequencing data, or raw accelerometer activity data, I wouldn't like to say. Just pulling out a few SNPs to work with should be fine, but a gwas might not.
When the level of credits was set, I believe it was intended to cover full costs for small projects. Based on this, I am fairly sure that there must have been several previous projects that completed in less than £1000.
If you are doing anything costly, you will need to become familiar with the Rate Card, and do tests to find the most efficient ways of working. See costs and billing and managing usage and storage costs .
Browsing data using the Cohort Browser does not incur costs. RStudio sessions are more costly than R sessions in Jupyterlab. If you start a job and expect it to be quite quick, but it goes on for a long time, consider terminating it in the Monitor tab.
You might like to add a comment to this similar question, to see if the original author of it has any advice.
Thank you for using the forum.
Hi Rachael,
Thank you for your answer and the signpost to the similar question. This is definitely helpful.
Please sign in to leave a comment.