Hello, I have a problem with online blastp. I am trying to perform a blastp search of preselected set of plant genes against joined Plants and Fungi databases in biopython. I managed to obtain results but I have noticed that every time i run the script, the return xml files vary in size. Sometimes the difference is minor, whereas it can also be huge (like 340kb vs 4mb). I am of course conscious that databases grow, but the situation concerns also running script three times in a row and gettimg different xml sizes
I thought that the problem is associated with operating system, so I forced my mac to be awake and on full performace all the time, also the computer is not used at all during the script run. Even with that conditions the results vary.
What can be a reason? And also do You recommend performing online blast or rather local blast after downloading databases?
Thank You in advance
if it is the same script and same input I would expect something goes wrong. I never used biopython to blast online before. But from the commandline the connection is refused sometimes. This happens if you blast to much and to fast in a short time. Did you checked your xml's? I think the error will be in there.
To solve it, you need to add a pause between sequences and catch the error. If the error comes up try the same seq again after lets say 5 sec. But check for errors first.
Timeout actually partly helped! I have obtained nearly identical results using 15 seconds timeout between searches. I have found out that the files that have wrong size contain a SIGXCPU message. Repeating blast search for the troublesome queries usually helps