Subcellular Localization

LocPro localization results for the two rubber dandelion protein FASTA sets, TK1151 and OH-1004; displayed data follow the input ID version.

Examples TkA01G624900 LXL04_000258

Dual-reference IDs

Input ID TkA01G481780

TK1151 ID TkA01G481780

OH-1004 ID LXL04_002123

Main locations use complete ESM2 + PROFEAT features and probability calibration against independent experimentally annotated reference proteins. This is not experimental localization in rubber dandelion. Multiple locations are supported, so probabilities need not sum to one. Detailed-location scores remain uncalibrated and are shown separately.

Calibration method and independent validation metrics

Search Results

Download calibrated main-location results

Top1 predicted location Nucleus

Top1 probability (benchmark calibrated) 0.425

Gene ID TkA01G481780

Data version TK1151

ModelLocPro

Model downloadLocPro model folder

PaperYintao Zhang et al. LocPro: A deep learning-based prediction of protein subcellular localization for promoting multi-directional pharmaceutical research. J Pharm Anal. 2025. doi:10.1016/j.jpha.2025.101255

Nucleus0.425

Rank 1 · TkA01G481780 · TK1151

Cell membrane0.251

Rank 2 · TkA01G481780 · TK1151

Endoplasmic reticulum0.247

Rank 3 · TkA01G481780 · TK1151

Cytoplasm0.219

Rank 4 · TkA01G481780 · TK1151

Plastid0.066

Rank 5 · TkA01G481780 · TK1151

Detailed locations (uncalibrated model scores)

These scores support exploratory ranking and must not be interpreted as localization probabilities. The model contains cross-species categories; these outputs do not establish the corresponding structures or localization in rubber dandelion.

Location	Raw score
Nucleus	0.422
Nucleus membrane	0.378
Endoplasmic reticulum membrane	0.305
Endoplasmic reticulum	0.301
Cytoplasm	0.243