Identify customized novel targets with Standigm ASKTM
Revolutionize your target identification process
Don’t be suffered by labor-intensive data collection.
Standigm ASKTM
Database provides you a comprehensive collection of state-of-the-art knowledge from literature and
public databases.
Stop the guesswork.
Standigm ASKTM provides you different
insights for novel targets with customizable machine learning approaches.
Standigm ASKTM provides customized modules for your target research.
According to your needs, we utilize diverse tools such as natural language technologies or
systems biology approaches.
If you want to experience Standim ASKTM,
please contact us at
[email protected]
and request for access to the trial version.
How Standigm ASKTM Does It
We provide novel target candidates that perfectly meet your needs.
Standigm ASKTM is an AI-aided drug target identification platform ready to be customized to your research context. It provides not only a prioritized target list for query disease/phenotype but also the evidence of prediction and the additional information surrounding the target-disease relationship.
Standigm Database
Comprehensive integration of diverse data sources followed by expert curation
Publication-derived information using Standigm's proprietary NLP technologies
Prioritization Algorithm
Harmonizing diverse AI approaches according to the context
- A knowledge graph-based deep learning model
- Natural language processing technologies
- A genotype-phenotype predictor based on genome-scale metabolic models
- A random-walk
model based on modified PPI network and omics data
Interactive User Interface
Served with interactive user interface
User-driven exploration of prioritization evidence Command-free database browsing
Look how Standigm ASKTM
revolutionize target identification journey
Targets encoded in the human genome
Diseases and abnormalities defined
in the Experimental Factor Ontology
Approved drugs and drugs
from clinical trials
25 types of known relations
between 8 types of nodes
Public curated source databases
Scientific literature containing
gene-disease associations
Insert or remove customer-defined information
Integrate context-specific information
Start from unbiased data provided by the customer
NLP technology
- Identify co-occurring patterns of certain word types
- Investigate information regarding specific disease-target pairs
MoA investigation
- Identify enriched pathways for designated gene sets
(e.g., downstream affected genes for a selected target)
Path analysis
- Identify important paths connecting the specified disease-target pairs
- Identify paths containing one or more designated genes
Expression analysis
- Identify expression patterns of designated gene sets in disease- specific data
(patient sample or cell lines)
Construct new modules in collaboration with the customer
Publications
Partnership FAQ
-
What does “novel target”
mean?
One thing to emphasize here is the importance of the target filtering required after targets are ranked. There are several layers to defining a novel target. Some people say a target is new when there is no research related to the target. Others think it is new when at most only a similar function is known. Therefore, we have various filtering templates to provide the level of target customers want. What we also want to emphasize is that target discovery doesn't take much time. Our AI models have already been built and are ready to use.
-
How long does it take to
select a novel target candidate?
It takes two (2) weeks to get a final target. During these two (2) weeks, we select target candidates using AI and obtain a final target through a filtering step.
-
Is it possible to
customize the database or integrate it with Standigm’s internal data?
Customization of the database is a general process for the "ASK platform". We first utilize public databases and up-to-date literature abstracts. Beyond these databases, the customer's own additional in-house databases can be built and, if desired, integrated into existing databases to be used as training data for our AI models.
-
What is the source of
your database?
All previously published biometric data is available. Any data that exists in pairs can be transformed into nodes and edges. Thus, you can utilize any database that contains data from pairs of genes, proteins, diseases, toxicities, protein interactions, expressi on levels, and more. Currently, the "ASK Platform" contains more than 38 databases. Please contact us if your are interested in the list of specific databases used in ASK Platform.
-
What is the network
relationship?
The network of the "ASK Platform" is called the "Knowledge Graph". In this graph, nodes and edges are a way to represent all data that exists in pairs. This includes all expressible relationships such as gene-disease, drug-disease, protein interactions, gene-gene expression levels, drug-toxicity, etc. This relationship information is extracted from various databases and current literature information. In the case of literature information, Standigm's NLP technology can be used.
-
How many nodes and edges
are involved?
The ASK Platform has three main nodes: disease, targets and compounds. Also, it contains the edges connecting these nodes described in relationship to each other. As such, we have more than 4 million edges and 70 thousand nodes in our knowledge graph.
-
Does the knowledge graph
build a different database for each project? Or does it build only one?
The knowledge graph includes all information in a single massive&heterogeneous graph. Since all information from various databases are connected in a single graph, various new sub-networks exist as an integrated form. So, the various projects utilize a single integrated knowledge graph. However, if you want to discover a target using your own in-house data, we can make a specific knowledge graph and train the AI model with it train it.
-
How does Standigm set
the causality of levels of graph?
We classify the relationship between disease and genes as a simple association and direct causation. Specifically, the causal information we use is experimentally proven data in terms of the occurrence of a disease caused by direct genetic modification. We utilize many databases that contain experimentally proven data.
