Patternmatch/specification.md

Specifications for the script

Use the Nontarget

• Clusters should have large enough number of fragments (3+ maybe)

• nontarget has a pattern match function.

• Check for chlorinated compounds (has n+2, n+4, ...)

• GPL License (good)

• Script to get output for all of the results

• tabbed-deliminated text file

• Which clusters contain chlorinated compounds, how many fragments, what other isotopes.

format

• Readable by R.
• Manipulatable by Tidyverse
• Include it in script?

Additional Notes

• Independent of the ordering (it might get corrupted)

• Just search for chlorinated/brominated compounds.

• Use intensity + relative abundances that isos shows.

• usually, use_charges will be 1

• Headers to use in script.

• Packages to do things in parallel (multiple threads)

• be able to configure the number of cores or threads used.

• make it a requirement to be tab-deliminated.

• eliminate clusters with only m/z.

Notes

• cutint represents high enough intensity.

• mztol is in terms of ppm.

• mzfrac is in terms of absolute.

• PPM = true always

• First pass - (identify all of the fragments where chlorinated/brominated compounds are found and carbon)

• If isotopes exist, which ones have chlorines/bromines...

• return a data frame.

• No statistical testing required (check if it is chlorine or bromine)

• Isotopes to check (Carbon, Sulfur, Chlorine, Bromine all have n+2) (check make.isos object)

• Tweak parameters (but also put them into a slideshow/record, see how good the results are).

• set all rules to false

Isotopes

• These isotopes will tend to have different intensities
• Part of the reasoning can be attributed to differences.
• Forseeable future, adapt to different isotopes.
• Capture multiple outputs of different isotopes.

Data Summary

• Flags for certainty of halogenated compounds

Test Notes

• mztol seems to have a big impact on chlorine/bromine positives.