The AI platform for drug discoveryDr.UG

AI - Powered

Drug Design

The DearDTI (Molecule Screening) and MolEQ (Lead Optimization) consisting of Dr.UG are DEARGEN's artificial intelligence technologies that specialize in designing drug candidates for targets. Through the Molecule transformer technology, the affinity score between a target and a small molecule can be predicted. Based on this, small molecule drug candidates with higher efficacy can be designed by ADMET (absorption, distribution, metabolism, excretion, toxicity) optimization.


DearTRANS is a platform that quickly compares and analyzes huge amounts of genomic data and presents the results visually.

Genomic Data Meta-Analysis

Numerous genome research data has been published, but there is a limit in using all of these data due to the batch effect caused by unknown bias in individual studies. DearTRANS provides services of integrating and analyzing data between disparate studies by solving this batch effect problem through meta-analysis technology. This can overcome the limitations of existing individual studies and dramatically increase the number of research data samples.

Differentially Expressed Gene(DEG) analysis

DearTRANS analyzes whether the average expression level of the same gene is significantly different under different conditions. It is a widely used method in bio, medical, and pharmaceutical fields. DearTRANS provides a user-friendly web service that makes DEG analysis and visualization easy without the use of complex tools or programming.

Gene Ontology(GO) Analysis

Gene ontology (GO) analysis is a statistical analysis method for assessing the correlation between gene sets and biological functions. GO analysis based on a database categorized by gene function can be used to interpret important biological functions and pathways. DearTRNAS searches for the functional profile of gene sets.

Visualization of Analysis Results

Once you’ve created a research hypothesis, you can prove and visualize it with just a click. According to your hypothesis, if you specified samples of genomic data analyzed through meta-analysis as an experimental group and a control group, DearTRANS will show results by comparing and analyzing DEG. DEG analysis results will be provided by visualizing into MA plot and Volcano plot. Besides, significant genes from the DEG results between the experimental and control group can be selected and displayed as a heat map and a bar plot. These visual aids can enhance the understanding of researchers and can be directly used in research papers and presentation materials.

Case Study ∙ Alzheimer's Disease

Based on the DearTRANS, DEARGEN discovered DDT3 (code name) that is a key transcription factor protein that is overexpressed in the brain tissue of Alzheimer's Disease patients. Also, through cross-validation using DearTRANS, it has confirmed that DDT3 regulates genes involved in the production of Aβ and tau proteins that are regarded as the cause of Alzheimer's disease.


WX presents biomarkers, based on raw transcriptome data. It is an artificial intelligence technology that can predict biomarkers, prognostic biomarkers and even the mode of action (MOA) of disease targets.

Target Discovery through Genome Wide Feature Data Analysis

WX presents highly accurate biomarkers by selecting a succinct set of genes that are most specific to disease from raw transcriptome data. Based on GWX (Gene WX) that is our own deep-learning algorithm, WX presents candidate genes that are the most likely to be chosen as biomarkers by recognizing differences in gene expression between a selected experimental and control group and calculating the scores of the importance. WX features highly accurate results increased by using the entire genome data instead of DEG data that is filtered by the difference in gene expression. By using the WX, target genes were currently selected from many pipelines such as Alzheimer ’s Disease, Rheumatism, Amyotrophic lateral sclerosis (ALS), and sarcopenia, etc.


Prognostic Biomarkers Prediction

Prognostic biomarkers are an important concept for diagnosing and treating disease. Predicting prognostic biomarkers enables medical treatments personalized to each patient's characteristics more in the strategy establishment for diagnosis and treatment of disease. We present a highly accurate prognostic biomarker by more accurately categorizing a high-risk and low-risk group among the relevant disease patients and calculating the importance of genome expression through Casexed Wx (CWx) that is own deep learning algorithm.


MOA (Mode Of Action) Prediction

In drug development, it is important to identify the mode of action (MOA). Drug development pipelines with an unclear MOA are not only costly in clinical trials but are also difficult to obtain approvals. DEARGEN predicts GO terms with a high priority by building Gene Ontology(GO) terms-based Neural Network Model and MOA by analyzing the weight of the Neural Network.

Case Study ∙ Lung Cancer

Two types of metastasis-induced biomarkers related to lung cancer prognosis have been found by using WX, and these biomarkers are shown to be significantly correlated to the prognosis through the analysis of actual clinical data. In a cell line experiment through the collaborative research with St. Mary's Hospital, it is confirmed that the metastasis of lung cancer cells has increased by miR-XXXa, b, which are two types of biomarkers related to lung cancer prognosis discovered by using WX. Also, the overall survival (OS) was analyzed by grouping according to the expression level of miR-XXXa, b in real patients' tissue samples. This result showed a significant correlation if the expression level of miR-XXXa, b is high, the prognosis worsens.

“For new drug design with our Dr.UG flatform, You can start Target Discovery here.”

Target Discovery Technology →