CREEDS: Crowd Extracted Expression of Differential Signatures

CREEDS (CRowd Extracted Expression of Differential Signatures) is a database of crowdsourced gene expression signatures for drug, genetic, and disease perturbations. The database was developed by the Ma’ayan Lab at the Icahn School of Medicine at Mount Sinai.

About CREEDS

CREEDS contains a collection of gene expression signatures extracted from the Gene Expression Omnibus (GEO). These signatures represent the differential expression of genes in response to:

• Drug Treatments: Gene expression changes after exposure to pharmaceutical compounds, capturing drug effects at the molecular level
• Gene Perturbations: Signatures from knockdown, knockout, or overexpression experiments showing how modifying specific genes affects overall gene expression
• Disease States: Differential expression between disease and normal tissue samples, highlighting the molecular basis of various pathological conditions

The signatures were extracted through a crowdsourcing approach, where contributors manually curated gene expression data from GEO to identify control and perturbation samples. This process involved:

1. Identifying relevant GEO datasets
2. Carefully selecting appropriate case/control samples
3. Processing raw data to generate differential expression signatures
4. Quality control and validation of the extracted signatures

Each signature consists of lists of up-regulated and down-regulated genes, along with metadata about the experimental conditions, sample characteristics, and statistical significance.

Data Access

The CREEDS database is freely accessible through a web interface at https://maayanlab.cloud/CREEDS/. Users can search for signatures by various criteria including drugs, genes, or diseases. The complete dataset is also available for download.

Applications

CREEDS signatures can be used for a variety of applications in biomedical research:

• Drug Signatures:
  • Drug repositioning and repurposing
  • Identifying mechanisms of action for compounds
  • Predicting drug side effects
  • Finding similar drugs with shared mechanisms
• Genetic Signatures:
  • Understanding gene function
  • Identifying gene regulatory networks
  • Pathway analysis and enrichment
  • Target discovery for therapeutics
• Disease Signatures:
  • Characterizing disease mechanisms
  • Identifying biomarkers
  • Drug discovery for specific diseases
  • Patient stratification and precision medicine

These signatures are particularly valuable when used for signature matching algorithms to connect drugs, genes, and diseases through their shared gene expression patterns.

Related Projects

CREEDS is related to other Ma’ayan Lab projects including:

• LINCS Program (Library of Integrated Network-Based Cellular Signatures)
• BD2K (Big Data to Knowledge)
• IDG (Illuminating the Druggable Genome)