Please Increase the limit of number of proteins to be analysed
I have got large datasets containing 257,295,361,460 and 504 proteins respectively. Its not possible to break them down in smaller sets for analysing with ProCKSI. Generally, even bigger datasets are reported in literature for analysis of protein structure and comparison. It would really be a great help particularly in my work to be able to analyse the large datasets with ProCKSI.
#1 Updated by Anonymous over 10 years ago
Currently, ProCKSI deals with each dataset as a whole for each task and needs to post-process the complete results in one go.
- Huge amounts of data have to be transferred from the compute node to the head node for post-processing.
- Standardising/clustering/etc. needs to deal with all results at once and can take too much time loosing the connection to the browser (see #39).
I have temporarily increased the number of proteins per dataset to 400.br
It seems that this is the utmost limit that ProCKSI can handle in the current design. br
The limit will possibly be descreased again as soon as more people want to use ProCKSI at the same time.
- Tasks have to be broken down into smaller junks. The size of the results will be smaller and much easier to handle.
- Post-processing tasks must be put into a (seperate, high-priority) queue and listed in the overview page similar to the main calculation tasks.