proteomics-ptm
Post-translational modification analysis including phosphorylation, acetylation, and ubiquitination. Site localization, motif analysis, and quantitative PTM analysis with MSstatsPTM.
SKILL.mdSkills/proteomics/proteomics-ptm/SKILL.md
GitHub🔬 Post-Translational Modification Analysis
Identify, quantify, and analyze post-translational modifications from mass spectrometry data.
Core Capabilities
- MaxQuant PTM output processing: Filter by localization probability, site annotation
- Site localization scoring: A-score calculation for confident PTM site assignment
- Motif analysis: Extract and count amino acid motifs around modification sites
- MSstatsPTM: Site-level quantification with protein-level normalization
- Differential PTM analysis: Adjusted for protein-level abundance changes
CLI Reference
python omicsclaw.py run proteomics-ptm --demo
python omicsclaw.py run proteomics-ptm --input <phospho_sites.txt> --output <dir>
Common PTMs and Mass Shifts
PTM_MASSES = {
'Phosphorylation': 79.966331, # STY
'Oxidation': 15.994915, # M
'Acetylation': 42.010565, # K, N-term
'Methylation': 14.015650, # KR
'Dimethylation': 28.031300, # KR
'Trimethylation': 42.046950, # K
'Ubiquitination': 114.042927, # K (GlyGly remnant)
'Deamidation': 0.984016, # NQ
'Carbamidomethyl': 57.021464, # C (fixed mod from IAA)
}
Algorithm / Methodology
Processing MaxQuant PTM Output
import pandas as pd
import numpy as np
# Phospho(STY)Sites.txt from MaxQuant
phospho = pd.read_csv('Phospho (STY)Sites.txt', sep='\t', low_memory=False)
# Filter valid sites
phospho = phospho[
(phospho['Reverse'] != '+') &
(phospho['Potential contaminant'] != '+')
]
# Filter by localization probability
phospho_confident = phospho[phospho['Localization prob'] >= 0.75]
print(f'Confident sites (prob >= 0.75): {len(phospho_confident)}')
# Extract site information
phospho_confident['site'] = phospho_confident.apply(
lambda r: f"{r['Gene names']}_{r['Amino acid']}{r['Position']}", axis=1
)
Site Localization Scoring
def calculate_ascore_simple(peak_matches_with_ptm, peak_matches_without_ptm, total_peaks):
'''Simplified A-score calculation'''
if peak_matches_without_ptm >= peak_matches_with_ptm:
return 0
p = peak_matches_with_ptm / total_peaks if total_peaks > 0 else 0
if p <= 0 or p >= 1:
return 0
from scipy.stats import binom
p_value = 1 - binom.cdf(peak_matches_with_ptm - 1, total_peaks, 0.5)
return -10 * np.log10(p_value) if p_value > 0 else 100
Motif Analysis
from collections import Counter
def extract_motifs(sites_df, sequence_col, position_col, window=7):
'''Extract sequence windows around modification sites'''
motifs = []
for _, row in sites_df.iterrows():
seq = row[sequence_col]
pos = row[position_col] - 1 # 0-indexed
start = max(0, pos - window)
end = min(len(seq), pos + window + 1)
# Pad if at sequence boundary
motif = '_' * (window - (pos - start)) + seq[start:end] + '_' * (window - (end - pos - 1))
motifs.append(motif)
return motifs
def count_amino_acids_by_position(motifs, center=7):
'''Count amino acid frequencies by position'''
position_counts = {i: Counter() for i in range(-center, center + 1)}
for motif in motifs:
for i, aa in enumerate(motif):
position_counts[i - center][aa] += 1
return position_counts
MSstatsPTM Site-Level Quantification (R)
library(MSstatsPTM)
# Prepare input from MaxQuant
ptm_input <- MaxQtoMSstatsPTMFormat(
evidence = read.table('evidence.txt', sep = '\t', header = TRUE),
annotation = read.csv('annotation.csv'),
fasta = 'uniprot_human.fasta',
mod_type = 'Phospho'
)
# Process data
processed_ptm <- dataSummarizationPTM(ptm_input, method = 'msstats')
# Differential PTM analysis (adjusting for protein-level changes)
ptm_results <- groupComparisonPTM(processed_ptm, contrast.matrix = comparison_matrix)
Parameters
| Parameter | Default | Description |
|---|---|---|
--method |
maxquant |
maxquant, msstatsptm, custom |
--mod-type |
Phospho |
PTM type to analyze |
--loc-prob |
0.75 |
Localization probability threshold |
--motif-window |
7 |
Motif window size |
Why This Exists
- Without it: Phosphorylation counts are meaningless if not normalized against total protein abundance
- With it: Decouples actual post-translational regulatory changes from sheer protein expression changes
- Why OmicsClaw: Streamlines complex localization scoring and statistical corrections (MSstatsPTM)
Workflow
- Calculate: Localise modifications using probabilistic scoring.
- Execute: Map sites to host proteins and correct for total abundance.
- Assess: Perform differential PTM abundance testing.
- Generate: Output motif logos and differential statistics.
- Report: Synthesize site-level Volcano plots and motif tables.
Example Queries
- "Run phosphoproteomics analysis on these MaxQuant sites"
- "Identify enriched motifs in hyper-phosphorylated peptides"
Output Structure
output_directory/
├── report.md
├── result.json
├── confident_sites.csv
├── figures/
│ ├── motif_logo.png
│ └── site_volcano.png
├── tables/
│ ├── differential_ptm.csv
│ └── motif_counts.csv
└── reproducibility/
├── commands.sh
├── requirements.txt
└── checksums.sha256
Safety
- Local-first: Strict offline processing without external upload.
- Disclaimer: Requires OmicsClaw reporting structures and disclaimers.
- Audit trail: Hyperparameters and operational flow states are logged fully.
Integration with Orchestrator
Trigger conditions:
- Automatically invoked dynamically based on tool metadata and user intent matching.
Chaining partners:
peptide-id— Upstream sequence identificationquantification— Upstream total protein normalization
Version Compatibility
Reference examples tested with: numpy 1.26+, pandas 2.2+, scipy 1.12+
Dependencies
Required: numpy, pandas, scipy Optional: MSstatsPTM (R), pyopenms
Citations
- MaxQuant — Cox & Mann, Nature Biotechnology 2008
- MSstatsPTM — Kohale et al., MCP 2023
- Phosphosite localization — Beausoleil et al., MCP 2006
Related Skills
peptide-id— Identify modified peptides upstreamquantification— Quantify PTM site intensitiesprot-enrichment— Pathway enrichment of modified proteins