TODO.md

In rickhelmus/patRoon: Workflows for Mass-Spectrometry Based Non-Target Analysis

Priority

Ext

• R bundle
  • steps
    • Create a shortcut/batch file?
      • use R.utils::createWindowsShortcut() to link Rgui.exe
• update GHPreRel ref in handbook
• cleanup
  • remove appveyor files/scripts from patRoon/patRoonDeps

Predict

• general
  • also score formulas?
    • general score mechanism currently not there
  • re-rank after addCompoundScore()
    • would make sense, but might break eg suspect annotation ranks
• quant
  • target RFs: those that are specified by user in a suspect list, to be used directly by calculateConcs()
  • filter suspects/annotations from conc results, ie to remove candidates with very low concs

General

• add showProgress option for future MP
• ppm spectral averaging (Ricardo)
• SIRIUS5: doc that tool should be present even if host is not a worker. Or another approach?
• No need for R.utils dep if normalizePath() is used everywhere instead of getAbsolutePath()?
• get rid of magrittr dependency?

Reporting

• get rid of pngquant
  • fully remove remainder when old reportHTML is fully removed (installation script, Dockerfile, options, ...)

Features

• Screening
  • sets: warn that re-grouping (adduct()<- / selectIons) will discard screening results
    • or somehow handle this? re-screen? although original suspects will be lost...
  • function that copies suspect adduct to annotations slot?
    • can only work with exactly one match
      • could throw a warning when this happens
    • tricky with sets
      • doc that screenSuspects need to be re-run?

TPs

• newProject()
  • selector for CTS transLibrary?

MS library

• prepareChemTable()
  • skip/warn if obabel is not available?

Low priority

General

• add 'keep.rownames = FALSE' to all as.data.table methods (or see if there is a work-around)
• remove mz column from patRoonData suspects?
• convertMSFiles()
  • Agilent .d is also a directory?
  • Remove necessity to have different input/output formats? (at least OK for pwiz)
• delete() for other classes
• generalize makeLegend() to new plot util

Features

• misc
  • OpenMS: alignment may yield negative RTs...
  • OpenMS MapAligner exception
    • seems to be related when little overlap between sets --> add note in doc?
• adduct annotations
  • selectIons(): prefer adducts based on MS/MS? eg handy for Na/K adducts
  • what to do with unsupported adducts for annotation?
    • default selectIons() to only consider 'common' adducts? or change default adducts for componentization algos?
    • check better for what is supported by SIRIUS?
• import XCMS features: verify anaInfo (or remove necessity, eg with importAnaInfo func)
• getEICsForFeatures method for kpic2?
• load OpenMS intensities in parallel
  • either with futures or with MP and cache intensities afterwards
• XCMS: multiple features grouped in same analysis?
  • can be, but now handled by default method="medret" param. Make this configurable?
• updatePICSet(): also sync peaks list? otherwise doc
• Somehow integrate XCMS::fillChromPeaks
• Normalization
  • maybe: allow usage of normalized intensities with filter()?
  • find another way to assign close/far ISTDs: if there are multiple close ones available, it makes more sense to not consider those that are a bit far away.

Annotation

• SusDat MF support
• parallel MSPeakLists generation?
• somehow handle different fragment formula annotations when making a consensus between formula/compounds objects
• DA formulas: also rank formula results like GF/SIRIUS?
• plotSpectrum/spectrumSimilarity: allow separate MSLevel for comparisons
• Support multiple MS/MS formula annotation candidates (ie same MS/MS peak annotated with different formulas)
  • mainly relevant for GenForm

Components

• feature components
  • cliqueMS
    • change checkPackage GH link once PRs are merged
    • maxCharge --> chargeMax (same as OpenMS)? update docs
• RC: check spearmans correlation
• NT: minimum size argument, combine rows for multiple rGroups?
• int and others: also use calculateComponentIntensities() for intensities?
• intclust
  • optionally take areas instead of intensities
  • cache results
• import check sessions?
  • needs way to match component names

tests

• checkComponents() / checkFeatures()
  • server tests?
  • import
• MSPeakLists and others?: also test object that is fully empty (now still has analyses)
• ensure peaklists are sorted
• features
  • new multiple blank filtering
  • syncing of XCMS/KPIC2 objects
  • check if featindex and groups slots are in sync with features
  • subsetting and groupScores
  • plotInt sets arg?
• components: somehow verify adductConflictsUsePref

MS library

• library
  • fixup annotation formulas too?
    • for now a little bit in C++ fixAnnFormula(), maybe implement verifyFormulas() in C++ someday?
  • future
    • plotSpectrum, plotVenn, same data format for MSPeakLists
• compounds
  • show mirror spectrum in report?
    • Would need library data somehow --> perhaps compoundsLibrary can include averaged lib spectra

Future

General

• test negative subset indices
• convertMSFiles(): Support OpenMS vendor conversion? (eg thermo)
• newProject()
  • also allow suspect annotation with only peak lists? currently only selectable if formulas/compounds selected
• Reduce non-exported class only methods
• future MP
  • delayBetweenProc?
  • batch mode
• msPurity integration
• algorithmObject() generic: for xset, xsa, rc, ...
• more withr wrapping? (par)
• newProject()
  • concentration column for anaInfo
  • generate more detailed script with e.g. commented examples of subsetting, extraction etc
  • import Bruker seq file?
  • fix multi line delete (when possible)

Features

• makeSet(): also support fGroups method via comparison?
• feature optim:
  • keep retcor_done?
  • get rid of getXCMSSet() calls?
• filter()
  • document which filters work on feature level (e.g. chromWidth)
  • remove zero values for maxReplicateIntRSD?
• integrate OpenMS feature scoring and isotopes and PPS in general (also include filters?)
• OpenMS: Support KD grouper?
• Integration of mzMine features (package pending...), MS-DIAL and peakonly?
• suspect screening
  • automatic suspect list name assignment if that's lacking? might be handy for some NORMAN lists
• topMost filter that accepts rGroups, either as AND or OR

Annotation

• MSPeakLists
  • isotope tagging is lost after averaging
  • test avg params
  • metadata() generic?
  • DA
    • generateMSPeakListsDA: find precursor masses with larger window
    • tests
      • utils? EICs with export/vdiffr?
      • test MS peak lists deisotoping?
• metadata for Bruker peaklists?
• SIRIUS: use --auto-charge instead of manually fixing charge of fragments (or not? conflicting docs on what it does)
• test score normalization?
• timeouts for SIRIUS?
• do something about negative H explained fragments by MF?
• MetFrag: auto-include suspect results if suspectListScore is selected?
• do something with sirius fingerprints? --> comparison?
• fix compoundViewer
• add new MF HD scorings and make sure default normalization equals that of MF web
• CFM-ID and MS-FINDER integration
• utility functions to make custom DBs for MetFrag and SIRIUS and support to use them with the latter
• DBE calculation for SIRIUS?
• OM reporting
• as.data.table: option to average per replicate group?
• ID levels for non-suspects
  • function to calculate ID levels from suspect list (to take RTs/MSMS if available), formulas, compounds
  • store in compounds?
  • does it make sense for formula candidates?
  • add into reporting
    • also mark if in suspect list

Suspects

• ID level rules: add scorings for SIRIUS/DA
• interface
  • also convert TASQ?
  • annotateSuspects()
    • check why it's is sometimes slow
      • seems to be logging, disable by default? --> only slow with testthat?
  • don't assign level <1 if suspect is a target? or give the choice (or make filter?)
• misc
  • prepareSuspectList(): export?
    • mainly to allow merging of lists, perhaps make util for that instead? Would also be handy for MF databases
      • could also fix column names, replace "-" with NAs etc
    • if yes, mention in ref docs for screenSuspects()

components

• mass defect components
• split peak correlation and adduct etc annotation? would allow better non-target integration
• fillPeaks for CAMERA (and RAMClustR?)
• feature components
  • cliqueMS
    • current adduct conversion to this format doesn't mimic Cat and 2H/2Na etc
      • Perhaps just document limitation?
  • minimal annotation abundance across analyses (eg adduct must be annotated in >=X analyses)?
  • prefAdducts: also include eg Na by default?

Sets

• compound/formula set consensus
  • weights for ranking (like compound consensus)?

TPs

• filter on stability/persistence/toxicity of TP?

rickhelmus/patRoon documentation built on April 25, 2024, 8:15 a.m.