Skip to content

Latest commit

 

History

History
129 lines (91 loc) · 7.62 KB

output.md

File metadata and controls

129 lines (91 loc) · 7.62 KB
Raw

lehtiolab/ddamsproteomics: Output

This document describes the output produced by the pipeline.

Output files

The output is a number of text, SQLite and HTML files. Depending somewhat on inputs, the following can be obtained from the output directory:

  • target/decoy PSM tables (TSV files)
  • a peptide table (TSV)
  • protein and genes tables (TSV)
  • quant and PSM SQLite lookup tables
  • a QC report (HTML)
  • a PTM PSM table (when searching PTM modifications)
  • a PTM peptide table (when searching PTM modifications)

File columns

The PSM tables are essentially an MSGF TSV table, with a number of extra fields for each PSM. By default the table is filtered on a PSM and peptide FDR of 0.01.

  • Retention time(min)
  • Ion injection time(ms)
  • Ion mobility(Vs/cm2)
  • missed_cleavage (count per PSM)
  • Master protein(s) (protein grouping using maximum parsimony is used for protein tables)
  • Protein group(s) content (Protein groups have members which can be explained by the representative protein)
  • Total number of matching proteins in group (Those proteins in the group that this PSM matches to)
  • Gene ID, Gene Name, Description (gene identity and protein description information from fasta)
  • percolator svm-score
  • PSM q-value and posterior error probability (PEP) (both calculated as TD competition by percolator)
  • peptide q-value and PEP (as PSM q-value but for peptides)
  • TD (target or decoy)
  • Biological set (sample or sample set name)
  • Strip (when fractionating samples you can supply an e.g. high-pH or HiRIEF strip name in --mzmldef)
  • Fraction (when fractionating samples, you can supply a fraction number/name)
  • MS1 area (summed MS1 intensity from Dinosaur detected feature aligned to a PSM)
  • FWHM (as MS1 intensity but the full width at half max of the feature)
  • ...plex channels (isobaric reporter intensity)
  • Experimental, Predicted and Delta pI (isoelectric point data for HiRIEF samples)

The peptides_table.txt is a merged multi-set peptide table derived from the PSMs. Peptide identifications are interpreted as the best scoring PSM. Then per sample (set) the following fields are given:

  • PSM count
  • q-value and PEP (same as PSM table peptide q-value for best PSM)
  • MS1 area (shows the highest area aka summed intenstiy of all filtered PSMs for the peptide)
  • ...plex channels, which are summarized isobaric data, described in the usage documentation
  • Fully quanted PSM count (for each peptide, how many PSMs without any missing isobaric value)
  • Quanted PSM count (how many PSMs with isobaric value in each channel)

Proteins and genes tables are like peptide tables, and contain similar fields. Of course there are differences:

  • Peptide count, Unique peptide count (the number of peptides for a protein/gene and number of peptides that uniquely match this protein group/gene)
  • q-value: genes are calculated using the picked FDR method from Savitski et al., 2015
  • MS1 precursor area (calculated using the top-3 highest intensity peptide for a protein/gene)
  • ...plex channels are as in peptide tables
  • logFC, count, sca.P.Value, sca.adj.pval are output from DEqMS analysis

When labile PTM modifications have been passed with --locptms, the pipeline runs luciphor2 to determine false localization rates for these PTMs, and also outputs a PTM-annotated PSM table. This contains the following extra fields:

  • Top luciphor PTM (best scoring peptide localization permutation)
  • Top PTM score (its score)
  • Top PTM FLR (its FLR)
  • High-scoring PTMs (other permutations with high scores)

Both --locptms and --ptms (for stabile modifications, e.g. Acetyl) result in a PTM-annotated peptide table. This reports the following differing fields

  • Master protein(s) (the master protein(s) matching this peptide, with annotation of PTM sites)
  • Protein(s) (other proteins matching)
  • Matching gene count
  • SAMPLE_SETNAME_PTM FLR (the FLR of the PTM-peptide in a sample set)
  • PTM flanking sequence(s) (the protein sequence flanking (+/- 7 aa) a PTM site, one for each site on a peptide

Note that when --totalproteomepsms is passed, the isobaric ratios in this table will be offset to the global search gene (default, proteins used if no genes exist) ratios, by subtracting those log2 ratio values. When also --normalize is used the medians of the totalproteome proteins (derived from that total proteome PSM table) are used for median-centering the PTM table.

PTM peptide tables can differ from normal peptide tables in a confusing way: when peptides are found in a sample/set, but PTM-peptides are NA in the same sample. This happens when there are multiple possible residues on which the PTM can be, and luciphor does not agree on the site with the search engine. The PTM peptide table will then contain NA for the search engine peptide, or the search engine peptide will be left out completely if it is not in any set according to luciphor.

Pipeline and tools

The pipeline is built using Nextflow and processes data using the following steps:

  • MSGF+ - Peptide identification search engine
  • Percolator - Target-decoy scoring
  • OpenMS - Quantification of isobaric tags
  • Dinosaur - Quantification of precursor peptides
  • Hardklor/Kronik - Quantification of precursor peptides
  • Luciphor2 - False localization rates for PTMs
  • Msstitch - Post processing, protein inference
  • DEqMS - Differential expression analysis

MSGF+

MSGF+ (aka MSGF+ or MSGFPlus) performs peptide identification by scoring MS/MS spectra against peptides derived from a protein sequence database. PMID 25358478

Percolator

Percolator is a semi-supervised machine learning program to better separate target vs decoy peptide scoring. PMID

OpenMS

OpenMS is a library that contains a large amount of tools for MS data analysis. This workflow uses its isobaric quantification program to extract peak intenstities for isobaric multiplex samples. PMID 27575624

Dinosaur

Dinosaur identifies peptide features in MS1 data and is an improved reimplementation of the MaxQuant algorithm. PMID 27224449

Hardklor/Kronik

Hardklor identifies peptide features in MS1 data by reducing isotope distributions to a single monoisotopic mass. It in tandem with its Kronik utility (which summarizes the Hardklor results from LC/MS data) can be used to quantify MS1 peptide features. This can be used instead of Dinosaur by passing --hardklor

Luciphor2

Luciphor2 is a site localization tool for generic post-translational modifications, and yields false localization rates for peptide PTM configurations. PMID 25429062

Msstitch

Msstitch is a software package to merge identification and quantification PSM data, reporting PSM, peptide, protein and gene tables, adding q-values, quantifications, protein groups, etc.

DEqMS

DEqMS is an R package for testing differential protein expression in quantitative proteomic analysis, built on top of the Limma package. PMID 32205417