This is related to issue <a class="issue-link js-issue-link" data-error-text="Failed t

Filter data on MS2 mz range,about rformassspectrometry/compounddb

Comments (2)

jorainer commented on August 16, 2024

Just some timings for fetching the data with and without filter.

The goal is to find MS2 spectra with a peak at an m/z of about 195.08. Below we use a CompDb with the full HMDB version 4.1.

library(CompoundDb)
cdb <- CompDb("CompDb.Hsapiens.HMDB.4.1.sqlite")

mz_in <- 195.08

## Get all spectra
system.time(all <- spectra(cdb))
   user  system elapsed 
 49.169   5.952  55.404 

## Get all spectra with peaks potentially in the same region
system.time(subs <- spectra(cdb, filter = ~ msms_mz_range_min <= mz_in & msms_mz_range_max >= mz_in))
   user  system elapsed 
 15.723   2.852  18.716 

length(all)
[1] 458901
length(subs)
[1] 187798

Still quite some spectra were retrieved, but the performance is better with the filter.

from compounddb.

jorainer commented on August 16, 2024

Now proceeding with finding the spectra that have an actual m/z matching 195.08 (using the hasMz method from #28).

system.time(all_has_pk <- hasMz(all, mz = mz_in, ppm = 20))
   user  system elapsed 
 10.930   0.203  11.212 

system.time(subs_has_pk <- hasMz(subs, mz = mz_in, ppm = 20))
   user  system elapsed 
  4.058   0.044   4.128 

sum(all_has_pk)
[1] 591

sum(sus_has_pk)
[1] 590

The same spectra were detected, with the exception of one, that was missed because the filter with msms_mz_range_min and msms_mz_range_max did not consider a ppm.

from compounddb.

Recommend Projects