CORTEX

Mechanism of Action

Understand how a treatment affects biological pathways and molecular processes. Cortex performs differential expression, pathway enrichment, transcription factor activity inference, and cross-condition comparison on your treated-vs-control omics data. The analysis produces a complete mechanistic picture: from individual gene changes through pathway-level effects to master regulator identification, with every finding grounded in statistical evidence.

What You Get

Deliverables

Affected pathways ranked by enrichment

GSEA enrichment across Hallmark, KEGG, and Reactome gene sets identifies which biological processes are activated or suppressed by treatment. For the IL-4/IL-13/IL-33 study, 38 significant pathways were identified: TNFα/NF-κB as the top IL-33 pathway (NES 2.27 in mast cells) and IL2/STAT5 as the top IL-4/IL-13 pathway (NES 2.01 in eosinophils). Cross-condition comparison reveals how different treatments engage different biological programs.

EosinophilsMast cells
TNFα/NF-κB (Eos IL-33)IL2/STAT5 (Eos IL-4)IL2/STAT5 (Eos IL-13)Allograft rej. (MC IL-4)Allograft rej. (MC IL-13)TNFα/NF-κB (MC IL-33)1.61.822.22.41.831.892.012.092.262.27

Dose-response patterns

Transcription factor activity inference (via decoupler with CollecTRI regulons) identifies the upstream regulators driving observed expression changes. STAT6 emerged as the master regulator in IL-4/IL-13-treated eosinophils (score 19.4/13.6), while NFKB1 dominated IL-33 responses in both cell types (score 52.3/53.3). MEIS2 was identified as a novel regulator, a finding that could not have been predicted from pathway analysis alone.

03006009001,2001,500IL-33IL-4IL-138506643761334646402EosinophilsMast cells

Cross-pathway interaction mapping

Systematic comparison across all treatment conditions reveals which effects are shared versus condition-specific. The key finding, that IL-33 drives the broadest response via NF-κB/AP-1 rather than STAT6, overturns the assumption that all type-2 cytokines signal through similar pathways. Over 90% of DEGs were cell-type-specific, demonstrating why cell-type resolution is essential for MoA studies.

STAT6NF-κBAP-1
activity scoreMC IL-4 / AP1Eos IL-13 / STAT6Eos IL-4 / STAT6Eos IL-33 / NFKBMC IL-33 / NFKB101020304050608.113.619.452.353.3
DECISION ENABLED

Identify on-target and off-target effects to inform lead optimization or safety assessment. Focus on treatments that engage intended pathways while monitoring for unexpected transcription factor activation or off-target pathway effects.

Sample Output

IL-4/IL-13/IL-33 mechanism of action in eosinophils and mast cells

GSE282860: IL-4/IL-13/IL-33 in Eosinophils and Mast Cells62 samples
4,272
TOTAL DEGs
62
SAMPLES
24 / 38
EOS / MC
DEG Hierarchy: Cytokine Response Strengthby cell type
DEGs03006009001,2001,500IL-33IL-4IL-138506643761334646402EosinophilsMast cells
CytokineEos (up/down)MC (up/down)
IL-33850 (617↑ 233↓)1334 (964↑ 370↓)
IL-4664 (481↑ 183↓)646 (411↑ 235↓)
IL-13376 (315↑ 61↓)402 (306↑ 96↓)
Hallmark Pathway Enrichment: Top PathwaysGSEA NES
PathwayEos IL-4Eos IL-13Eos IL-33MC IL-4MC IL-13MC IL-33
TNFα/NF-κBn/an/a1.83n/an/a2.27
IL2/STAT51.892.01n/an/an/an/a
Allograft rejectionn/an/an/a2.092.26n/a
Transcription Factor Activities: Master Regulatorsdecoupler inference
Activity scoreMC IL-33NFKB1MC IL-4AP1Eos IL-33NFKBEos IL-13STAT6Eos IL-4STAT6010203040506053.38.152.313.619.4
STAT6 NF-κB AP-1
Key Findings14 findings identified
IL-33 drives the broadest transcriptional response via NF-κB/AP-1, not STAT6
Over 90% of DEGs are cell-type-specific
MEIS2 identified as a novel regulator
IL-13 produces zero TF activity in mast cells
IL-33 upregulates IL13 (log₂FC +10.0) and IL5 (+7.4) in mast cells

FLUOXETINE NSCLC MECHANISM (GSE200209)

Transcription Factor Ranking: SREBF1 vs ATF4#1 vs #3 of 688 TFs
Activity difference#3 ATF4#2 SREBF2#1 SREBF100.30.60.91.21.51.80.123 (n.s.)1.563 *1.723 *

ATF4 (published mechanism) is NOT significant (padj=0.456). SREBF1 is the true driver (#1/688, diff=1.723, padj=0.024). ATF4 mRNA is also unchanged (log2FC=-0.108, padj=0.665). This overturns the published mechanism.

Fluoxetine: Key Pathway Statscholesterol vs ER stress vs MYC
2.936
CHOLESTEROL NES
0.963
ER STRESS NES (n.s.)
-3.615
MYC_TARGETS_V1 NES
10:1
UP/DOWN RATIO

MYC paradox: MYC mRNA unchanged but MYC_TARGETS_V1 NES=-3.615, suggesting post-translational suppression. ER stress not significant (NES=0.963, FDR=0.574). Cholesterol biosynthesis is the dominant program (NES=2.936).

Fluoxetine: Top DEGs (151 up, 15 down)166 DEGs total
Genelog2FCPathway
HMGCS1+2.62Cholesterol biosynthesis
INSIG1+2.37Cholesterol homeostasis
FADS2+2.23Fatty acid desaturation
MSMO1+2.08Sterol biosynthesis
SCD+1.81Lipid desaturation

ATG7 ALZHEIMER'S DOUBLE VULNERABILITY (GSE286094-GSE286095)

ATG7: Double Vulnerability Axesanti-correlated (r=-0.130, p<10^-30)
UPR impairment (ATF6)Ferroptosis (GSEA NES)-2-10122.5-1.27+1.94

UPR impairment (ATF6 -1.27) and ferroptosis activation (NES +1.94) are anti-correlated: ATG7-knockout microglia face simultaneous protein quality control collapse and oxidative vulnerability. 8 growth/survival pathways collapsed, with only p53 and TNFa remaining active (DAM state).

ATG7: Key Pathway Changesbulk DEGs at 20 months
PathwayNESDirection
Ribosome+2.89UP
Translation+2.71UP
IFN response+2.34UP
Protein processing in ER-2.64DOWN
Antigen presentation-2.31DOWN
N-glycan biosynthesis-2.22DOWN
How It Works

Methodology

STEP 1

DESeq2 differential expression per condition

Each treatment condition is analyzed independently using DESeq2 with donor-level covariates. For the 62-sample dataset (24 eosinophil, 38 mast cell), this produced condition-specific DEG lists: from IL-33's 1,334 mast cell DEGs down to IL-13's 376 eosinophil DEGs.

STEP 2

GSEA pathway enrichment across gene set collections

Gene Set Enrichment Analysis runs across Hallmark, KEGG, and Reactome collections for each condition. The analysis identified 38 significant pathways with condition-specific enrichment patterns, revealing that IL-33 and IL-4/IL-13 engage fundamentally different biological programs.

STEP 3

Transcription factor activity inference

Decoupler with CollecTRI regulons infers upstream transcription factor activities from the observed expression changes. This identified STAT6, NFKB1, AP-1, and the novel regulator MEIS2 as the key drivers of the observed transcriptional responses.

STEP 4

Gene regulatory network construction

Condition-specific regulatory edges are constructed from TF-target relationships, producing networks that show how master regulators propagate their signals through downstream gene programs.

STEP 5

Cross-condition concordance metrics

Systematic comparison across all conditions and cell types quantifies shared versus unique effects. Concordance metrics (Pearson, Spearman, Jaccard) measure the degree of overlap, revealing that over 90% of DEGs are cell-type-specific.

Who This Is For

Target personas

Pharmacologist

Understand treatment mechanisms at the pathway and transcription factor level to inform pharmacological models.

MoA scientist

Get a complete mechanistic picture from gene-level changes through master regulator identification.

Lead optimization team

Compare MoA profiles across compounds or conditions to guide optimization toward desired biological effects.