Stanford School of Medicine

LaneConnex
Lane Medical Library & Knowledge Management Center

 

a division of IRT

[searching]

What is the LOCATE database?

What is it?

LOCATE is a web-accessible, curated database of predicted protein properties derived from the conceptual translation of mouse transcripts:
  1. computational predictions of protein membrane organization
  2. subcellular localizations
Furthermore, the subcellular locations of selected proteins have been determined by a high-throughput immunofluorescence-based assay, and by the manual review of over 1700 peer-reviewed publications.

What is it for?

LOCATE stores classifications of predicted protein organization based on predicted signal peptide and transmembrane domain(s). In some cases, predictions have been validated experimentally.

LOCATE is useful for determining:
  1. the predicted membrane topology of a protein
  2. the predicted subcellular localization of a protein
By virtue of the homology principle, homologous and paralogous proteins may be cautiously inferred to share these (predicted) properties.

What is in it?

LOCATE classifies proteins into one of five categories of membrane organization:
  1. soluble intracellular protein (no transmembrane domains or signal peptide)
  2. soluble secreted protein (signal peptide, no transmembrane domains)
  3. type I membrane protein (one transmembrane domain, signal peptide)
  4. type II membrane protein (one transmembrane domain, no signal peptide)
  5. multi-pass membrane protein (multiple transmembrane domains)
The mouse proteome dataset used in LOCATE is the IPS7 FANTOM3 Isoform Protein Sequence set generated by the RIKEN FANTOM consortium.

This dataset is comprised of protein sequences conceptually translated from mouse full-length transcripts generated by FANTOM. Transcript sequences are clustered into transcriptional units (TUs), wherein a TU is a grouping of transcripts that arise from a single genomic locus and share at least one nucleotide having the same genomic location and orientation.

What is the process?

Translations of TUs are processed by a high-throughput, automated pipeline, which combines publicly available feature predictors with empirically determined annotation rules. Protein orientation with respect to the membrane is achieved by merging the predictions of several well-characterized protein secondary structure predictors:

Key references

Source

Lane Librarian

Record created 9/15/2007.

ypouliot, September 18, 2009

Click here to access LaneConnex without logging in for full access to free resources. Licensed content restricted. Click here to log in for full access to all LaneConnex resources.