Meaning of n_db and n_ptm values

[katherine-d21]

Hello!

I am using PepQuery 2.0.2 and am looking for more detail on how to interpret the 'n_db' and 'n_ptm' columns in the psm_rank.txt file.

When these columns contain -1, 0, or 1 - what do these mean in terms of passing or failing validation?

Thank you in advance,
Katherine

[wenbostar]

Hi @katherine-d21 , please find the explanation about both n_ptm and n_db at http://pepquery.org/document.html#saoutput. I just updated the documentation.

[katherine-d21]

Hello @wenbostar - Thank you for your reply and for updating the documentation!

I have two follow-up questions:

1. What are the parameters/values that are considered when assigning a yes/no for confidence of a PSM? I attempted to look through the code, and from my understanding, the p-val =< 0.01, rank =1, n_db = 0, n_ptm = 0, appear to be the criteria for assigning 'yes' for a confident PSM. Is this correct?

2. What happens if an input peptide has a length that falls outside of the peptide length parameters (min and max lengths)? For example, if I have a set of input peptides that vary in length from 7 to 17mer, but am only interested in those that are 8-10mer long. So I set PepQuery min and max lengths to be 8 and 10, respectively. How does PepQuery validate (or not validate) peptides that are longer than 10mer (in this example)?

Thank you in advance,
Katherine

[wenbostar]

1. What are the parameters/values that are considered when assigning a yes/no for confidence of a PSM? I attempted to look through the code, and from my understanding, the p-val =< 0.01, rank =1, n_db = 0, n_ptm = 0, appear to be the criteria for assigning 'yes' for a confident PSM. Is this correct?

Yes, except the default threshold for p-value filtering is 0.01 for peptides longer than 8 and 0.05 for peptides with length equal or shorter than 8. This is described in the "PSM scoring" section of the PepQuery2 paper: https://www.nature.com/articles/s41467-023-37462-4.

if I have a set of input peptides that vary in length from 7 to 17mer, but am only interested in those that are 8-10mer long.

I'm a bit confused about this. If you're only interested in 8-10mer peptides, could you just filter your input peptides to only keep peptides with 8-10mer for PepQuery analysis?

[katherine-d21]

Hello @wenbostar ,

1. Thanks for clarifying the confident PSM criteria!

2. My example oversimplified my issue, so I will share more details to help inform your response:

I am analyzing non-tryptic peptides from immunopeptidomics data, using the "Non enzyme" option. The one sample I'm using in my testing yielded peptides from 7 to 17mer in length.

Due to high memory requirements of no enzyme and immunopeptidomics analysis, my workaround has been to heavily restrict the peptide length parameters in PepQuery to enable the analysis to finish within a reasonable amount of time. Using two MGFs (from my one test sample) as my input MS/MS data, I have found that considering one peptide length (where the minLength value is equal to that of maxLength) greatly reduces PepQuery runtime.

However, restricting the peptide length (in each PepQuery run) means that I have to run PepQuery multiple times, each with a different peptide length, to analyze one sample's peptides. Thus, for my one test sample, I would run PepQuery eleven times to analyze peptides from 7 to 17mer in length.

Another caveat is that the 7 to 17mer range is specific to the one sample used in my testing (i.e., a different sample could yield peptides of 8 to 16mer, and another sample could have 9 to 18mer peptides, and so on).

After discussing with senior lab members, there were two possible approaches mentioned:

• One approach was similar to your suggestion: step 1) restricting min/max lengths in PepQuery (e.g., minLength = maxLength = 7, to only consider 7mer peptides) and step 2) filtering the peptides before using them in PepQuery (e.g., select only 7mer peptides from original peptides as input). Repeat this for each length found in the input peptides. For my one test sample, I'd prepare eleven subsets of peptides (7-17mer) as input for eleven PepQuery runs (7-17mer).

• The second approach was to use the input peptides as they are (no filtering by length before PepQuery). So I'd use the same set of input peptides for every PepQuery run, each run set up with a different lengths. For my one sample, I'd have one input set of peptides for eleven PepQuery runs (7 to 17mer).

The second approach was what I was asking about in my previous comment: by using the (unfiltered) peptides as input for every PepQuery run, some will inevitably fall outside of PepQuery's minLength and maxLength parameters, and thereby could affect the filtering steps and pass/fail validation.

I am having difficulty choosing which approach is more suitable. I hope this explanation gives more information, and I would greatly appreciate any suggestions/advice you can share, especially if there's a better solution.

Thank you again in advance,
Katherine