Test consistency of data input

[sneumann]

Hi, to ensure we have consistent input, we need unit tests for

1. Reading the showcase
2. Reading from three input files (Height, MSP, Sirius)
   and use the showcase numbers as gold standard, unless we can explain the difference and consider that superior.

MetFamily (Production, R 3.6.3, 1.3.2-metfamily-79bd66f5fd-nv4c8)

Load Example Data, filter Intensity: 10000, fold-change: 2
Number of filtered MS1 features: 219 / 2585

Import Data, Metabolite_profile_showcase.txt and MSMS_library_showcase.msp

The data import was successful.
2640 / 5824 spectra were imported successfully. (15 empty, 3163 low intensity, 5 too heavy)
2414 / 2640 spectra were successfully mapped to MS¹ features. (226 unmapped)
34369 / 145973 fragments were successfully imported. (1807 too heavy, 109797 low intensity)
2414 / 5823 MS¹ features were successfully imported. (420 were isotopes, 2989 without spectra)

Load Example Data, filter Intensity: 10000, fold-change: 2
Number of filtered MS1 features: 209 / 2414

MetFamily branch devel, commit 0a349b9d988c0b30e67ae0500d32831b851ebc1e

R version 4.3.3 (2024-02-29), MetFamily build: 1.3.5-IPB-6335, MetFamily package: 0.99.5

An error occurred while reading the input files. Please check the file format and content and try again. The error was:
Error in h(simpleError(msg, call)): error in evaluating the argument 'x' in selecting a method for function 'rowData': (converted from warning) restarting interrupted promise evaluation 

=> Missing fixes by Norman ?

Import Data, Metabolite_profile_showcase.txt and MSMS_library_showcase.msp, canopusShort.txt

The data import was successful.

2640 / 5824 spectra were imported successfully. (15 empty, 3163 low intensity, 5 too heavy)
2435 / 2640 spectra were successfully mapped to MS¹ features. (205 unmapped)
34369 / 145973 fragments were successfully imported. (1807 too heavy, 109797 low intensity)
2435 / 5823 MS¹ features were successfully imported. (373 were isotopes, 3015 without spectra)

But when I do my standard filter Intensity: 10000, fold-change: 2
then I get Number of filtered MS1 features: 110 / 2435
=> That seems far too low compared to the above 209 / 2414

Yours, Steffen

[sneumann]

New test with 762a45f (gpatoine-read_qfeat)
Import Data, two files: Metabolite_profile_showcase.txt and MSMS_library_showcase.msp

2640 / 5824 spectra were imported successfully. (15 empty, 3163 low intensity, 5 too heavy)
2414 / 2640 spectra were successfully mapped to MS¹ features. (226 unmapped)
34369 / 145973 fragments were successfully imported. (1807 too heavy, 109797 low intensity)
2414 / 5823 MS¹ features were successfully imported. (420 were isotopes, 2989 without spectra)

Load Example Data, filter Intensity: 10000, fold-change: 2
Number of filtered MS1 features: 209 / 2414

=> Same as MetFamily (Production, R 3.6.3, 1.3.2-metfamily-79bd66f5fd-nv4c8) above.

[sneumann]

The off-by-ten (209 vs. 219) is acceptable, and not traceable anymore. 209 is reproducible now.
We should re-do the showcase data, adding SIRIUS along, and create the data for "Load example data" through the vignette for reproducibility.