bigbio / pIR

Estimation of isoelectric point of peptides and proteins using Support vector machines
10 stars 6 forks source link

Performance #3

Closed ypriverol closed 8 years ago

ypriverol commented 8 years ago

@enriquea We should try to work around the performance issues. It would be great if this 60000 peptides can run in less time, please review the Bjel methods steps of the algorithm and also in the iterative algorithm.

enriquea commented 8 years ago

Hi @ypriverol Both Bjell and Iterative method were changed to improve the performance. You could check the changes done here: https://github.com/enriquea/pIR. You could test it running the following commands:

branca <- read.table(file = "data/ruiBrancaDataSet.csv", header = TRUE, sep = ",") data <- mdply(branca, function(sequence, pIExp) { pIBjell(sequence = sequence, pkSetMethod = "expasy") })

In the present version the performance was improved 5X.

Let me know how it works for you,

Best regards

ypriverol commented 8 years ago

@enriquea Can we test the SVM to see how it works.

ypriverol commented 8 years ago

the closed pull-request solve all the performance issues. https://github.com/ypriverol/pIR/pull/10