Weird numbers in precursor_error_distribution.csv

[microdou]

I'm trying to understand the data shown in reports\precursor_error_distribution.csv.

Here are the questions:

• First, what does each column mean? The data table is presented without column titles. It looks like the data is sorted by column 3 in descending order. I figured out columns 1 & 2 are Precursor_Mass_Error(Da) & Precursor_Mass_Error(ppm), but I couldn't figure out what columns 3-5 mean.

• Second, some numbers appear to be too large... Numbers in column 3 abruptly become as large as e+038 towards the end of the table (as shown below), which can be resulted from some calculation bug. If that's the case, I'm concerned that this bug might filter out some positive identifications.

-0.01869,   -8.54541, -39.5518,       0, 3
-0.041835, -17.1084,  -39.7004,       1, 3
-0.041589, -10.5107,  -39.89,         1, 3
-0.033696, -10.0624,  -39.9388,       2, 3
 0.030256,  16.337,   -40.204,        1, 3
-0.024363, -14.2231,  -41.0635,       0, 3
 0.06527,   17.4623,   -3.40282e+038, 1, 3
-0.046203, -18.2091,   -3.40282e+038, 1, 3
-0.037315, -14.5916,   -3.40282e+038, 1, 3
-0.027676, -16.8244,   -3.40282e+038, 1, 3
-0.047786, -18.7007,   -3.40282e+038, 2, 3
-0.037989, -11.4851,   -3.40282e+038, 1, 3
-0.036913, -18.3455,   -3.40282e+038, 0, 3

[01joy]

Dear @microdou ,

The file reports\precursor_error_distribution.csv contains precursor erors for all PSMs, and the file reports\filtered_precursor_error_distribution.csv contains only the precursor errors for all filtered PSMs. Both files share the same column meanings. They are: Precursor_Mass_Error(Da), Precursor_Mass_Error(ppm), SVM score, Target or Decoy, and Peptide Type. You can find the meaning of these columns in the wiki: https://github.com/pFindStudio/pLink2/wiki/CSV-result.

The SVM score e+038 just means this PSM is poorly matched, it has lost the meaning as a score. Those PSMs with score e+038 are most decoy PSMs and have not passed the FDR filter.