Executive Summary
LRX-1 is NOT an LRP family protein - the automated annotations are incorrect.
Analysis Results
1. Cysteine Pattern Analysis
- Total cysteines: 30 (8.1% of sequence)
- Cysteine spacing pattern: [13, 8, 8, 5, 6, 8, 19, 7, 7, 4, 11, 9, 7, 5, 14, 22, 8, 5, 8, 6, 7, 8, 5, 9, 7, 8, 7, 9, 7]
- Canonical LDL-A domains found: 0
- Why no LDL-A domains?: The cysteine spacings (e.g., 13, 8, 8, 5, 6) are incompatible with the canonical LDL-A pattern which requires specific spacing like C-x(2,3)-C-x(3,4)-[DE]-x(4,5)-C-x(7,8)-C-x(2,3)-C-x(3,9)-C
2. Cysteine-Rich Regions
Identified 17 distinct cysteine-containing regions:
- Region 1: positions 151-159 (YAINYCDKR)
- Region 2: positions 202-210 (TSCSHVFFQ)
- Region 3: positions 211-221 (CSIGQTFPLA)
- And 14 more regions throughout the protein
Key finding: All 30 cysteines are in the C-terminal 2/3 of the protein (none in first 100 aa)
3. Membrane Topology Analysis
- Signal peptide (1-19): Weak - only 40% hydrophobic residues
- Transmembrane helices: None detected (no regions with >70% hydrophobicity over 20 aa)
- Predicted topology: SECRETED protein, not membrane-bound
4. LRP Family Comparison
| Feature |
True LRP Proteins |
LRX-1 |
Match? |
| Size |
>4000 aa |
369 aa |
❌ |
| LDL-A domains |
30-40 |
0 |
❌ |
| β-propeller domains |
Yes |
No |
❌ |
| EGF-like domains |
Yes |
No |
❌ |
| Membrane protein |
Yes |
No |
❌ |
5. Domain Predictions Validation
UniProt Claims vs Reality:
- "4 LDL receptor Class A domains" → FALSE: No canonical LDL-A patterns
- "Single-pass type I membrane protein" → FALSE: No transmembrane helix
- "Endomembrane system" → FALSE: Likely secreted
- "Vesicle-mediated transport" → UNSUPPORTED: Based on false LRP assignment
6. What LRX-1 Actually Is:
- A small (369 aa) cysteine-rich protein
- Likely secreted, not membrane-bound
- Contains novel, non-canonical cysteine-rich domains
- Possibly C. elegans-specific protein family
- Function remains unknown
Conclusions
- The name "LRX-1" (LRP X-hybridizing) is misleading - this is not an LRP family member
- UniProt's automated ARBA predictions are wrong - demonstrates importance of validation
- The protein needs reclassification - should not be annotated as LRP-related
- GO annotations should be removed - all are based on incorrect domain predictions
Recommendations
- Remove all current GO annotations (they're based on false premises)
- Reclassify as "uncharacterized cysteine-rich protein"
- Experimental work needed to determine actual function
- Check if this represents a nematode-specific protein family
Code Used
See lrx-1-bioinformatics-analysis.py for the complete analysis code.