henhaobio
Sign in

Scientific DataCell BiologyPeer-reviewed journal article

The Signaling Pathways Project: an integrated ‘omics knowledgebase for mammalian cellular signaling pathways

Scott A. OchsneriD, David Abraham, Kirt A. Martin, Wei DingiD, Apollo McOwitiiD, Wasula H. KankanamgeiD, Zichen WangiD, Kaitlyn J. Andreano, Ross A. HamiltoniD, Yue CheniD, Angelica Hamilton, Marin L. GantneriD, Michael DehartiD, Shijing Qu, Susan Galloway HilsenbeckiD, Lauren B. BecneliD, Dave BridgesiD, Avi Ma’ayaniD, Janice M. Huss, Fabio StossiiD, Charles E. FouldsiD, Anastasia Kralli, Donald Patrick McDonnelliD, Neil J. McKennaiD

PDFDOIPreprintData
In-depth full-text digest available · Jump to digest ↓

Abstract

Integrated mining of public transcriptomic and ChIP-Seq datasets has the potential to illuminate facets of mammalian cellular signaling pathways not yet explored in the research literature. Here, we designed a web knowledgebase, the Signaling Pathways Project (SPP), which incorporates stable community classifications of the four major categories of signaling pathway node (receptors, enzymes, transcription factors and co-nodes) and their cognate bioactive small molecules (BSMs). We then mapped over 10,000 public transcriptomic or cistromic experiments to their relevant signaling pathway node, BSM or biosample of study. To provide for prediction of pathway node-target transcriptional regulatory relationships, we generated consensus omics signatures, or consensomes, based on measures of significant differential expression of genomic targets across all underlying transcriptomic experiments. To expose the SPP knowledgebase to researchers, a web browser interface accommodates a variety of routine data mining strategies. Consensomes were validated using alignment with literature-based knowledge, gene target-level integration of transcriptomic and ChIP-Seq data points, and in bench experiments that confirmed previously uncharacterized node-gene target regulatory relationships. SPP is freely accessible at https://beta.signalingpathways.org.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 3 (3 Jun 2019) · CC BY 4.0

Signaling Pathways Project maps over 10,000 transcriptomic and ChIP-Seq experiments to signaling nodes and ranks node targets by meta-analysis (consensomes), validated against literature and bench assays.

Most data points in public transcriptomic and ChIP-Seq datasets go unexamined in their source articles, and their limited findability and re-use hampers modeling of signaling pathways. This resource pools any genetic or small-molecule perturbation of a node family into a node-centric meta-analysis (prior approaches were perturbation-centric), integrates ChIP-Seq evidence, and serves results through a curated web resource for researchers without informatics training.

Key results

  1. ShownThe Signaling Pathways Project (SPP) organizes public omics experiments under community classifications of receptors, enzymes, transcription factors and co-nodes plus bioactive small molecules (BSMs), served through a public web knowledgebase.

    Families: 174 receptor, 616 enzyme and 371 transcription factor (from IUPHAR, IUBMB and TFClass); over 10,000 transcriptomic or ChIP-Seq experiments mapped; typically 20-25% of archived transcriptomic datasets failed a three-step QC filter; dataset DOIs minted through CrossRef. · Abstract; Data model design; Transcriptomic datasets; Signaling Pathways Project web application

  2. ShownTop-ranked targets in nuclear receptor transcriptomic consensomes mostly match targets already known from the literature, and they also rank high in the matching ChIP-Seq consensomes.

    Top ten targets from each of ERs-Hs-MG and AR-Hs-Prostate (human) and GR-Mm-Liver and PPARs-Mm-Metabolic (mouse): 36/40 (90%) previously reported as node-family targets; 33/40 (82%) at or above the 90th percentile in comparable ChIP-Seq consensomes. · Canonical signaling node downstream targets are…

  3. ShownThe most frequently regulated targets in a pan-node mouse liver consensome are enriched for metabolic enzyme genes, especially those linked to human metabolic disease and those catalyzing rate-limiting steps.

    Top 1% of All nodes-Mm-liver (n = 258): 38% (99/258) encode metabolic enzymes vs 5.7% (1490/25922) consensome-wide, a 6.5-fold enrichment; human orthologs of 42% (41/97) enzyme targets have OMIM metabolic-disorder deficiencies vs 15% (3865/25922); 40/97 (41%) rate-limiting vs a published 14% (96/687). · Frequently regulated hepatic transcripts are…

  4. ShownBench assays confirmed hormone regulation and growth phenotypes for two high-ranking targets whose roles in estrogen or androgen receptor signaling were unexplored: TPD52L1 for estrogen receptors in MCF-7 cells and MBOAT2 for the androgen receptor in LNCaP cells.

    TPD52L1: ERs-Hs-All-TC CPV <1E-130 (99.99th percentile), ChIP-Seq 99th percentile; 17BE2 induction in MCF-7 blocked by 100 nM FULV; siRNA reduced 17BE2-induced proliferation. MBOAT2: AR-Hs-All TC CPV = 3.7E-38 (99.96th percentile); induced by 0.1 nM R1881 in LNCaP; knockdown raised day 5 cell numbers with R1881 only. · Bench validation use case 1: elevated consensome…; Figure legends

  5. ShownIntersecting consensomes nominated targets that were then confirmed in mouse models: Ppp1r3c and Prkab2 as glycogen synthase regulators where GR, ERR and insulin receptor signaling converge, and the uncharacterized Mcrip2 as an ERR/Ppargc1 target in muscle cells and brown adipocytes.

    Ppp1r3c GR-Mm-All CPV = 1.89E-15, regulated by 250 nM DEX in Hepa-1-c cells; Prkab2 ERR-Mm-All CPV = 1.84E-04, reduced by 40% in Esrra-/- muscle (p < 0.05, n=4); Mcrip2 ERRs-Mm-All CPV = 1.54E-12, Esrra and Ppargc1a occupied its intronic ERR element by ChIP, and Mcrip2 fell in ERR-deficient brown adipocytes. · Bench validation use case 2: GR, ERR family…; Bench validation use case 3: the murine ERR…; Figure legends

Methods, data and reproducibility

Biocurators mapped GEO transcriptomic and ChIP-Atlas-processed SRA ChIP-Seq experiments to nodes, bioactive small molecules and biosamples. Transcriptomic consensomes rank genes by a binomial consensus p-value (CPV) counting experiments with p ≤ 0.05 per node family, biosample and species, ignoring direction; ChIP-Seq consensomes average MACS2 peak values.

  • Biocurated GEO transcriptomic datasets (expression arrays, RNA-Seq) with genetic or bioactive small molecule perturbation of signaling nodes · human, mouse, rat · n = over 10,000 experiments (transcriptomic and ChIP-Seq combined)
  • ChIP-Atlas-processed SRA ChIP-Seq experiments annotated by immunoprecipitated node, BSMs and biosample
  • Murine metabolic enzyme gene set from the Mammalian Metabolic Enzyme Database · mouse · n = 1647
  • OMIM genes whose deficiency is linked to a human genetic disorder or trait · human · n = 5277
  • MCF-7 (ER, TPD52L1) and LNCaP (AR, MBOAT2) cell lines for bench validation · human
  • Mouse Hepa-1-c hepatoma cells (GR, Ppp1r3c) and C2C12 myoblasts/myotubes (ERR, Prkab2, Mcrip2) · mouse · n = Prkab2/Ppp1r3c C2C12 assays: three trials, 3 replicates per trial · ATCC CRL-1772 (C2C12)
  • Skeletal muscle from 12-week-old wild-type and Esrra-/- littermate mice · mouse · n = n=4 (Prkab2 Q-PCR)
  • Primary brown adipocytes from floxed-ERR mice, recombined with CRE lentivirus or overexpressing Ppargc1a/Gadd45g · mouse
Methods & tools
Ominer query interface and Regulation Reports (SPP), Bioconductor differential expression pipeline, ChIP-Atlas (uniform MACS2 peak calling), IUPHAR/Guide To Pharmacology, TFClass, IUBMB Enzyme Committee classification, OMIM, Mammalian Metabolic Enzyme Database, Oracle 12c, Java EE 6, JavaServer Faces/PrimeFaces, D3.JS, RT-qPCR and ChIP-qPCR, siRNA knockdown, adenoviral expression and luciferase reporter assays, CellTiter-Glo viability assay; GraphPad PRISM
Reagents & environment
Java Enterprise Edition 6 application on an Oracle 12c database; UI built with JavaServer Faces, PrimeFaces and D3.JS; supported browsers: Firefox 24+, Chrome 30+, Safari 5.1.9+, Internet Explorer 9+. Bioconductor pipeline version unstated.
Reproducibility gaps
Source code and bulk site download were waived at submission, with release promised upon publication; Bioconductor pipeline details, software versions and consensome scripts are not given in the text.
Limitations (6)
  • Coverage is limited to transcriptomic and ChIP-Seq data; other omics modalities such as metabolomics are deferred until such datasets become more numerous and better stewarded. · stated by the authors
  • Biases of publicly archived datasets toward particular nodes and biosamples are, to some extent, reflected in SPP. · stated by the authors
  • Targets regulated only in narrow contexts (for example acute ligand dosing or a single organ) can rank low when matching experiments are missing, so a low consensome rank does not exclude regulation. · stated by the authors
  • Because consensomes are mRNA-based, they miss node effects at the protein level, such as altered protein stability or translation rate. · stated by the authors
  • The CPV binomial test treats each contrast as an independent trial, yet all possible contrasts per dataset were computed and often share samples; correlated contrasts may inflate significance, so extreme CPVs (e.g. <1E-130) are better read as ranks than calibrated p-values. · noted in this summary
  • Validation reports proportions and fold-enrichments without formal significance tests, low-ranked control gene sets, or comparison against alternative ranking schemes such as fold-change or effect-size meta-analysis. · noted in this summary

Read this if you want to mine public transcriptomic and ChIP-Seq evidence for which signaling receptors, enzymes or transcription factors regulate your gene, or need a node-centric meta-analysis design.

Citations
122

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISLiaskos '26Figueiredo '21Domingo‐Fer '19Piran '20Vaparanta '24Chen '19
53% semantic matchSystems Biology
David Liaskos, Omid Oftadeh, Margherita Tonini+2 more·bioRxiv·Jun 22, 2026