Broad Institute LINCS Resources


The Broad Institute was funded as an U54 Data Generation and Analysis center for LINCS 2011-2014 and charged with creating a first installment of a large-scale transcriptional dataset.

LINCS resources created during this grant period are freely available for all users. These resources include:

  1. Landmark Genes. We developed a new approach to expression profiling based on a reduced representation of the human transcriptome. Namely, we identified 1,000 transcripts from which the remainder of the transcriptome can be computationally inferred, and we measure these 1,000 'Landmark' transcripts on Luminex beads. The list of landmark genes is available (xlsx, excel file).
  2. L1000 Standard Operation Procedure (SOP).The full protocol is available (PDF, n=50 pages)
  3. Cell Treatment Protocols. Cell culture and treatment protocols tailored to cell types tested in LINCS (PDF, n=13 pages).
  4. Data. L1000 data, numbering over 700,000 experiments along with experimental meta data are freely available for download via FTP from Username: "cmap-lincs" and Password: "l1000" (no quotes). Note (a) please use the log-in credentials provided, not anonymous FTP (b) given the data size, it might take time to download
  5. Code. Signature-processing code including the current inference model are available in the l1ktools Github repository.
  6. Signatures and LINCS Connectivities. Tools to look-up LINCS signatures and their connections.
  7. Meta data and inventory list Annotations on perturbagens profiled and a list of all data downloadable (xlsx, excel file)
  8. Mapping to probe pool, detection set and plate names Plate names are derived by combining a series name (e.g CPC), a particular plate number (e.g 005) and details on detection set (e.g DUO, UNI, bead batch). To derive the precise mapping please use worksheet #4 in the excel file ("lincs_lxb_id to broad file name") which will allow you to map a given LXB file to the plate, probe pool and detection set unambiguously. Mapping between detection set and probe pool is provided in worksheet #3 ("detection sets").
  9. Inference model The model currently in use at the Broad, which is trained on ~12,000 samples from GEO, is provided pre-publication and contains the inference model weights .mat file (MATLAB)
  10. Using FTP to download With over 800,000 files, listing the contents by FTP will be very slow. Instead please download the inventory above and then wget every individual file
  11. Support.Please email questions to
  12. Usage terms and conditions. Before using these resources, please review the license.
  13. Note.Please also see (described below)


In the next phase of LINCS, we plan to develop perturbational profiles for LINCS using L1000, explore sequencing-based approaches for transcriptional profiling of a wider set of genes and profiling novel genetic reagents including improved gain-of-function (cDNA) and loss-of-function (CRISPR) genetic libraries

As the NIH and our aims take shape, we will post more information. In the interim, if you have any questions, please contact us at

Other Resources

  • Connectivity Map Version 1

    A database of ~7,000 Affymetrix gene expression profiles representing ~1,300 compounds. This resource continues to be made available in maintenance-mode.


    We have developed a modern, cloud-based informatics platform that is engineered to support the dramatically large datasets that the L1000 assay has enabled. Academic users would probably find access via more useful as it includes bench biologist friendly user interfaces, visualization methods and collaboration tools. (it is a superset of the resources on this website).

  • Towards a reference database

    Our mission in Connectivity Map is to develop comprehensive signatures of cellular states and the tools to analyze them. Despite generous NIH and philanthropic contributions over the past 10 years, there still remains much to be done to execute on the original vision. We are developing ideas to support this effort. Stay tuned for more information.