lincs-l1000

is a Data Source.

The LINCS L1000 is a high-throughput, reduced representation gene expression profiling assay developed as part of the NIH Library of Integrated Network-Based Cellular Signatures (LINCS) Program. L1000 directly measures 978 landmark genes and computationally infers the expression of 11,350 additional genes, enabling cost-effective large-scale transcriptional profiling at approximately $2 per sample. The technology was developed to create a Connectivity Map (CMap) that catalogs cellular responses to genetic and chemical perturbations, facilitating drug discovery, mechanism of action determination, and functional annotation of genetic variants. The LINCS L1000 dataset comprises over 1.3 million gene expression profiles representing responses to 19,811 chemical compounds, genetic perturbations targeting 5,075 genes (via shRNA knockdowns and cDNA overexpression), and 314 biologics across multiple cell lines and time points.

Domains

drug discovery

License

Warning: No license entered

Homepage

lincs-l1000

Repository

GitHub

Infores ID

Unknown

FAIRsharing ID

Unknown

Product Summary

Products

From this Resource
ID Name URL Category Format Description
lincs-l1000.cmap LINCS Connectivity Map (CMap) clue.io GraphicalInterface http The Connectivity Map (CMap) database ...
lincs-l1000.clue CLUE Platform clue.io GraphicalInterface http The CLUE platform provides interactiv...
lincs-l1000.geo ⚠ LINCS L1000 GEO Dataset acc.cgi?acc=GSE92742 Product http LINCS L1000 data deposited in the Gen...
From other Resources
ID Name URL Category Format Description
spoke.graph SPOKE Graph ❔ GraphProduct ❔ The SPOKE knowledge graph containing ...
alzkb.browser AlzKB Graph Database Browser login GraphicalInterface http A browser interface for a knowledge g...
alzkb.data AlzKB Data Release (Version 2.0.0) v2.0.0 GraphProduct ❔ Memgraph data release for AlzKB.

Details

Overview

The LINCS L1000 platform represents a major advance in high-throughput transcriptional profiling technology. By measuring only 978 carefully selected “landmark” genes and using computational inference for the remaining transcriptome, L1000 achieves dramatic cost reduction (approximately $2 per sample) while maintaining high reproducibility and comparability to RNA-seq. This breakthrough enabled the generation of over 1.3 million gene expression profiles as part of the NIH LINCS Program, creating one of the largest publicly available transcriptional profiling resources.

Key Features

Technology

  • Landmark genes: 978 directly measured genes selected through data-driven analysis of 12,000+ microarray profiles
  • Inferred genes: 11,350 genes computationally inferred with 81% showing high accuracy
  • Platform: Ligation-mediated amplification (LMA) with Luminex bead-based detection
  • Cost: Approximately $2 per profile (vs. traditional microarrays or RNA-seq)
  • Reproducibility: >88% of technical replicates show Spearman correlation >0.9
  • Validation: Highly comparable to RNA-seq (median cross-platform correlation 0.84)

Dataset Composition (CMap-L1000v1)

  • 1,319,138 L1000 profiles comprising:
    • 19,811 small molecule compounds (drugs, tool compounds, screening libraries)
    • 18,493 shRNAs targeting 5,075 genes for loss-of-function studies
    • 3,462 cDNAs for gain-of-function studies
    • 314 biologics
    • 473,647 signatures (consolidating biological replicates)
  • Cell line coverage:
    • 9 core cancer cell lines (Touchstone reference dataset)
    • Up to 77 cell lines for Discovery dataset
    • Includes neuronal cell types (NPCs and differentiated neurons)
  • Time points: 6h and 24h for chemical perturbations, 96h for genetic perturbations

Applications

  • Drug discovery: Mechanism of action (MOA) determination, off-target effect identification
  • Compound annotation: Functional classification of uncharacterized molecules via connectivity to Perturbagen Classes (PCLs)
  • Genetic variant interpretation: Assess functional impact of disease-associated alleles (e.g., FBXW7, KEAP1, PTEN)
  • Clinical trial analysis: Evaluate target engagement and identify resistance mechanisms
  • Connectivity mapping: Discover relationships between genes, drugs, and disease states through gene expression signatures
  • Bioactivity screening: Identify transcriptionally active compounds from screening libraries

Data Access and Tools

Primary Data Portals

  • CLUE Platform (https://clue.io): Interactive analysis tools, signature search, data downloads, and APIs
  • GEO (GSE92742): Raw and processed data at multiple preprocessing levels
  • GitHub (https://github.com/cmap/cmapM): Pre-processing code and tools

Additional LINCS Resources

  • SigCom LINCS (https://maayanlab.cloud/sigcom-lincs): Search across 1.5M+ signatures
  • L1000FWD (https://maayanlab.cloud/L1000FWD/): L1000 Characteristic Direction Signature Search Engine
  • iLINCS (http://www.ilincs.org): Integrated LINCS data portal
  • LINCS Data Portal (http://lincsportal.ccs.miami.edu/dcic-portal/): DCIC data coordination center

Data Levels

  1. Level 1: Raw bead count and fluorescence intensity
  2. Level 2: Deconvoluted data (assigning expression to two genes per bead color)
  3. Level 3: Normalized data (LISS + quantile normalization) with inferred gene expression
  4. Level 4: Differential expression (z-scores)
  5. Level 5: Replicate-consensus signatures

Methodology Highlights

Perturbagen Classes (PCLs)

171 high-confidence Perturbagen Classes have been defined, representing groups of perturbagens with shared mechanisms. These include:

  • 92 compound classes (e.g., HDAC inhibitors, kinase inhibitors)
  • 60 loss-of-function gene classes
  • 17 gain-of-function gene classes

PCLs enhance interpretability by aggregating similar perturbagens to strengthen on-target signals while diminishing off-target effects.

Consensus Gene Signatures (CGS)

To mitigate strong off-target effects of shRNAs (where seed sequence effects often exceed on-target effects), a Consensus Gene Signature algorithm was developed that identifies consistent gene expression changes across multiple shRNAs targeting the same gene.

Query Methodology

The Connectivity Map uses a weighted connectivity score (WTCS) approach similar to Gene Set Enrichment Analysis, computing similarity between query gene signatures and database signatures. Results are normalized and summarized across cell lines using quantile-based metrics (τ scores) to identify robust connections.

Key Publications and Citations

Primary Reference:

  • Subramanian A, Narayan R, Corsello SM, et al. A Next Generation Connectivity Map: L1000 platform and the first 1,000,000 profiles. Cell. 2017;171(6):1437-1452.e17. doi:10.1016/j.cell.2017.10.049

Additional Resources:

  • Original Connectivity Map concept: Lamb J, et al. The Connectivity Map: using gene-expression signatures to connect small molecules, genes, and disease. Science. 2006;313(5795):1929-1935.
  • Detailed protocols: https://clue.io/sop-L1000.pdf
  • Connectopedia knowledge base: https://clue.io/connectopedia

Funding

Supported by NIH grants including 5U54HG006093, U54HG008699, and 5U01HG008699 as part of the NIH Common Fund LINCS Program.

Is this information incorrect or incomplete? Request an update.

Created: January 08, 2025 | Last modified: November 08, 2025