A Reference Library for Characterizing Protein Subcellular Localizations by Image-Based Machine Learning

Libraries composed of 789,011 and 523,319 optically validated reference confocal micrographs of 17 and 20 EGFP fusion proteins localized at key cell organelles as landmarks in murine and human cells were generated for assignment of subcellular localization in mammalian cells. For each image of individual cells, 160 morphology and statistical features were used to train a random forests classifier to automatically assign the localization of proteins and dyes in both cell types and to analyze the sequence requirements that specify subcellular localization of a model tail-anchor (TA) protein. Automated assignment of subcellular localizations for a library of TA proteins with randomly mutated C-terminal targeting sequences enabled the discovery of motifs responsible for targeting TA proteins to mitochondria, endoplasmic reticulum and the late secretory pathway. Analysis of directed mutants enabled refinement of subcellular localization motifs that characterize cellular sub-compartments.

Schormann W, Hariharan S, Andrews DW

Browse Data in IDR

idr0072-schormann-subcellref

idr0072-schormann-subcellref/screenA

idr0072-schormann-subcellref/screenB

Download

Data is available for download via Globus: idr0072-schormann-subcellref.

To download individual files in your browser, you can browse original data.

Download as JSON.

For more download options, including FTP, see the IDR Data download page.