DrugReflector

Rank candidate compounds from a transcriptional signature.

This UI is for research screening. You provide either an .h5ad matrix or a prepared v-score signature, and the model returns compounds ranked by predicted relevance.

这是一个科研筛药界面。你提供转录组 signature,系统会返回更值得优先验证的候选化合物排序。

Purpose Prioritize compounds for follow-up experiments
Input V-scores or H5AD expression matrix
Output Rank, logit, and probability for each compound
Open API Docs

Quick Start

Three practical steps for a first run.

01

Prepare a signature

Use either a v-score list like TP53,1.2 or a preprocessed .h5ad file.

02

Run ranking

Choose the input mode below, set Top N, and submit the job.

03

Export results

Download CSV or JSON and use the ranking as a short list for follow-up experiments.

Service Status
Checking model status...
Model Library
Loading...
Expected Use
Signature-driven compound prioritization

Run a Prediction

Choose the input mode that matches your data.

One gene per line. Use either GENE,score or GENE score.

5 genes detected in the current text box.

Best for preprocessed data. This model is most reliable when the uploaded matrix already represents a usable signature or a compatible expression input.

Results

Run a prediction to populate the table.

No predictions yet. Submit a signature or H5AD file to see ranked compounds.