Thermal Proteomics for Mechanism of Action (MoA)

Label-free, proteome-wide thermal profiling that reveals not just which proteins your compound binds, but how those interactions cascade through the cellular network.

Unlike affinity-based methods that report only direct physical binding and require compound modification, thermal proteomics captures the full proteome-wide response to compound treatment — revealing direct targets, downstream pathway perturbations, and polypharmacology in a single experiment.

Key Advantages:

Label-free — no compound modification or immobilization required
Proteome-wide — 5,000–8,000 proteins quantified per experiment
Direct + indirect effects distinguished in one assay
Works in live cells, lysates, and tissue samples
Integrates with phosphoproteomics and transcriptomics

Plan Your MoA Study View sample requirements

Thermal proteomics for MoA showing compound binding, thermal stabilization, and melting curve comparison.

Why MoA MoA Info Workflow Comparison Applications Sample Demo Case Study FAQ

Why Thermal Proteomics for MoA? Beyond Target ID

When a phenotypic screening campaign delivers a hit compound with a clear cellular effect but an unknown mechanism, the discovery team faces a familiar bottleneck: which protein is the compound actually engaging, and is that engagement the cause of the observed phenotype?

Thermal proteomics answers both questions in a single label-free experiment. Unlike affinity-based methods that only report direct physical binding (and require compound modification), thermal proteomics captures the full proteome-wide response to compound treatment — revealing not just which proteins are directly bound, but which pathways shift downstream as a consequence. This distinction matters because a compound that stabilizes its target and also destabilizes downstream signaling proteins tells a different mechanism story than a compound that only hits its target directly.

Our thermal proteomics for MoA service is designed around this interpretive advantage. We combine temperature-gradient thermal profiling with quantitative DIA or TMTpro mass spectrometry to generate melting curves for 5,000–8,000 proteins per experiment, then apply a structured interpretation framework that classifies each thermally shifted protein as a direct target, a downstream responder, or a polypharmacology hit.

For related capabilities, see our Thermal Proteome Profiling (TPP) service for a broader technology overview.

What MoA Information Thermal Proteomics Reveals

Thermal proteomics generates three distinct layers of mechanism information from a single experiment. Understanding these layers is essential for interpreting the data correctly and extracting maximum value from your study.

Layer 1: Direct Target Binding

When a compound binds directly to a protein, it alters that protein's thermal stability — typically shifting its melting temperature upward (stabilization) or, less commonly, downward (destabilization). These direct-binding events produce the largest thermal shifts in the dataset and are the most reproducible across biological replicates.

The key interpretive signal: a direct target shows a thermal shift in both live-cell and lysate-based formats. Because lysates eliminate cellular physiology, any shift observed in lysates must arise from direct physical binding, not from downstream signaling.

Layer 2: Downstream Pathway Perturbation

Proteins that are not directly bound by the compound can still show thermal shifts as a consequence of pathway engagement. For example, a kinase inhibitor that blocks its target's activity may cause downstream substrate proteins to change phosphorylation state, which in turn alters their thermal stability.

These downstream shifts are valuable mechanistic information — they tell you which pathways are functionally engaged by the compound. The interpretive key: downstream shifts appear in live-cell experiments but disappear in lysate-based experiments, because the signaling context is absent.

Layer 3: Polypharmacology and Off-Target Profile

Many drug candidates engage multiple proteins beyond their intended target. Thermal proteomics captures this polypharmacology comprehensively, revealing every protein whose thermal stability changes upon compound treatment.

This is particularly valuable for assessing selectivity risk early in discovery, identifying on-target toxicity mechanisms, discovering beneficial off-target activities for drug repurposing, and characterizing PROTACs and molecular glues that engage the degradation machinery and target proteins simultaneously.

Thermal Shift Pattern Interpretation

Thermal Shift Pattern	Cell-Based Result	Lysate-Based Result	Interpretation
Strong stabilization	✓	✓	Direct target binding
Strong stabilization	✓	✗	Downstream pathway effect
Weak stabilization	✓	✓	Low-affinity direct binder
Destabilization	✓	✓	Direct binding with conformational change
Destabilization	✓	✗	Signaling-dependent destabilization
Multiple proteins shifted	✓	✓ (subset)	Polypharmacology (direct binders + downstream)

Our Thermal Proteomics MoA Workflow: From Compound to Mechanism Story

Our workflow is designed to generate interpretable MoA data from the first experiment. Each step is optimized for data quality and mechanistic clarity.

Experimental Design Consultation

We begin by reviewing your compound properties, target hypotheses (if any), sample type, and study goals. Together we select the appropriate thermal profiling format: standard temperature-range TPP for discovery, 2D-TPP (temperature × concentration) for potency ranking, or PISA for higher-throughput screening. We also plan the orthogonal validation strategy — typically including thermal stabilization assay (PISA-based) or LiP-MS for hit confirmation.

Sample Preparation and Compound Treatment

Your samples (live cells, lysates, or tissue homogenates) are treated with the compound at the selected concentration(s) and incubation time. Vehicle controls and positive controls are included in every experiment. For 2D-TPP experiments, we use 5–8 concentrations spanning 3–4 log units around the expected active range.

Thermal Challenge and Fractionation

Samples are heated across a temperature gradient (typically 37–67°C, 8–10 temperature points) using a programmable thermal cycler. After heating, soluble (non-aggregated) protein fractions are separated from precipitated protein by centrifugation. The soluble fraction contains proteins that remained folded at each temperature.

LC-MS/MS Acquisition

Soluble fractions are digested and labeled (TMTpro for multiplexed TPP) or analyzed label-free (DIA for PISA). We use Orbitrap Eclipse or timsTOF Pro platforms with optimized gradients to maximize proteome coverage. Each temperature point is analyzed in triplicate.

Thermal Shift Analysis and Curve Fitting

Raw MS data is processed using our bioinformatics pipeline (Spectronaut, DIA-NN, or custom R/Python scripts). Melting curves are fitted for each protein using a sigmoidal (Boltzmann) model. The melting temperature (Tm) shift between treated and control conditions is calculated for every quantified protein.

MoA Interpretation and Reporting

This is the critical step that distinguishes our service. We apply a structured interpretation framework:

Direct target candidates: Proteins with significant Tm shifts in both cell and lysate formats
Downstream responders: Proteins with Tm shifts only in the cell-based format
Polypharmacology hits: Multiple direct targets identified across protein families
Pathway enrichment: GO/KEGG/Reactome enrichment analysis of all shifted proteins

The final report includes: complete protein quantification matrix, melting curves for all shifted proteins, ranked target candidate list with confidence scores, pathway enrichment analysis, and a written MoA summary with data interpretation.

Six-step vertical workflow for thermal proteomics MoA: design, treatment, thermal challenge, MS acquisition, analysis, and interpretation.

For complementary approaches, see our Thermal Shift Assay-MS (TSA-MS) service for target engagement confirmation.

How Thermal Proteomics Compares with Other MoA Approaches

Choosing the right method for MoA elucidation depends on your specific question, compound type, and available resources. The table below compares thermal proteomics with the most common alternatives.

Dimension	Thermal Proteomics (TPP/PISA)	ABPP (Activity-Based)	Pull-Down / MS	Phosphoproteomics	CRISPR Screens
Compound modification required	No	Yes (probe design)	Yes (immobilization)	No	No
Proteome coverage	5,000–8,000 proteins	Target family-specific	Bait-dependent	10,000+ phosphosites	Genome-wide
Direct targets detected	Yes	Yes	Yes	No	No
Downstream effects detected	Yes	No	No	Yes	Yes
Native cellular context	Live cells or lysates	Live cells or lysates	Lysates only	Live cells	Live cells
Throughput	Medium (10–100 compounds)	High	Low	Medium	Low
Label-free	Yes	No	No	Yes	Yes
Time to result	4–6 weeks	6–10 weeks	4–8 weeks	4–6 weeks	3–6 months

Selection Strategy: Choose thermal proteomics when: (1) your compound cannot be modified (natural products, clinical-stage compounds), (2) you need both target ID and downstream pathway context from a single experiment, (3) you are working with multiple analogs and need to compare their proteome-wide engagement profiles, or (4) you suspect polypharmacology and want comprehensive off-target assessment.

When to Use Thermal Proteomics for MoA: Key Application Scenarios

SCENARIO 1

Phenotypic Hit Deconvolution

Your phenotypic screen identified a compound that produces a desired cellular effect, but the molecular target is unknown. Thermal proteomics provides an unbiased, proteome-wide list of candidate targets. By comparing cell-based and lysate-based thermal profiles, we can distinguish the primary target from downstream effectors, giving you a focused list for validation.

SCENARIO 2

Off-Target Safety Assessment

A lead compound shows efficacy but raises safety concerns. Thermal proteomics profiles the compound against the entire expressed proteome, identifying off-targets that may explain toxicity. This information guides medicinal chemistry efforts to improve selectivity and informs early de-risking decisions.

SCENARIO 3

PROTAC Degradation Profiling

PROTACs engage both a target protein and an E3 ligase, then induce target degradation. Thermal proteomics captures this complex mechanism: the target protein shows decreased thermal stability (due to ternary complex formation), while the E3 ligase components may show stabilization. The proteome-wide view also reveals whether the PROTAC degrades off-target proteins.

SCENARIO 4

Natural Product MoA Elucidation

Natural products often cannot be modified for affinity-based methods. Thermal proteomics requires no compound modification, making it ideal for natural product MoA studies. The label-free nature preserves the compound's native binding properties.

SCENARIO 5

Analog Comparison for Candidate Selection

When comparing multiple analogs from a medicinal chemistry campaign, thermal proteomics reveals which compounds have the cleanest engagement profiles. Compounds that engage only the intended target (and perhaps a few well-understood off-targets) can be prioritized over those with broad polypharmacology.

For related dose-response capabilities, see our dose-response thermal profiling service.

Sample Requirements and Project Planning

Sample Type	Minimum Quantity (per condition)	Recommended Quantity	Format	Notes
Adherent cell pellets	5 × 10⁶ cells	1 × 10⁷ cells	Dry pellet, snap-frozen	Wash with PBS before freezing
Suspension cell pellets	5 × 10⁶ cells	1 × 10⁷ cells	Dry pellet, snap-frozen	Remove culture medium completely
Tissue (fresh frozen)	20 mg	50–100 mg	Snap-frozen in cryovial	Avoid freeze-thaw cycles
Tissue (FFPE)	5 × 10 µm sections	10 × 10 µm sections	Unstained slides	Deparaffinization performed in-house
Cell lysates	500 µg protein	1–2 mg protein	In compatible lysis buffer	Avoid high urea or detergents
Plasma/serum	50 µL	100 µL	Clear aliquot, no hemolysis	Depletion recommended for deep coverage

Important planning considerations:

Controls: Include vehicle-only control (DMSO or appropriate solvent) and a positive control compound with known targets
Replicates: Minimum 3 biological replicates per condition; 2–3 technical replicates per temperature point
Temperature points: 8–10 points for standard TPP; 5–8 concentrations × 6–8 temperatures for 2D-TPP
Turnaround: Typical project completes in 4–6 weeks from sample receipt

Representative MoA Data: What Thermal Proteomics Results Look Like

Thermal proteomics generates rich, multi-dimensional data. Below are the key result types you can expect from a typical MoA study.

Melting Curve Shifts

The primary data output is a set of melting curves comparing treated and control samples. For each quantified protein, we generate a sigmoidal melting curve with calculated Tm values. Direct targets show clear Tm shifts (typically 2–8°C), while downstream responders show smaller but reproducible shifts.

Proteome-Wide Thermal Shift Volcano Plot

The volcano plot visualizes the entire proteome's response to compound treatment. Each point represents a protein, plotted by its log₂ fold change in thermal stability (X-axis) against statistical significance (Y-axis). Proteins are color-coded by interpretation category: direct targets (red), downstream responders (blue), polypharmacology hits (orange), and no significant change (gray).

Pathway Enrichment from Thermal Hits

Pathway enrichment analysis of all thermally shifted proteins reveals the biological processes affected by compound treatment. This is particularly valuable for distinguishing on-target pharmacology from off-target effects — if the enriched pathways are consistent with the intended mechanism, the compound is acting as expected; if unexpected pathways appear, further investigation is warranted.

Dose-Response Thermal Profiles (2D-TPP)

For 2D-TPP experiments, we generate dose-response curves for each shifted protein, with EC₅₀ values that reflect the potency of target engagement. This allows direct comparison of compound potency across multiple targets and reveals whether off-target engagement occurs at concentrations relevant to the intended pharmacology.

Case Study: Large-Scale MoA Characterization of 166 Compounds Using Proteome-Wide Thermal Shift Assays

Van Vranken JG, Li J, Mintseris J, et al. "Large-scale characterization of drug mechanism of action using proteome-wide thermal shift assays." eLife, 2024. https://doi.org/10.7554/eLife.95595 [CC BY 4.0]

Background

Understanding how small molecules engage the proteome is essential for determining mechanism of action, but most methods are limited to single-target or affinity-based approaches that require compound modification. The field needed a scalable, label-free method capable of characterizing compound MoA across hundreds of compounds simultaneously.

Methods

Van Vranken et al. developed a high-throughput PISA workflow and applied it to 96 compounds with known mechanisms of action in live K562 cells, then further assayed 70 of these compounds in native K562 lysates. Each experiment quantified approximately 6,800 proteins per treatment in cells and approximately 7,840 proteins per treatment in lysates, generating over 1.5 million compound-protein thermal stability measurements in total.

Results

The cell-based PISA assay detected significant thermal shifts for 71% (56/79) of compound-target pairs where the target was quantifiable. In lysates, 72% (43/60) of pairs showed significant shifts. Combining both formats increased detection to 82% (49/60). The screen revealed extensive off-target engagement — Palbociclib was found to bind PLK1 kinase, AZD-7762 showed promiscuous kinase engagement including FER, LYN, CSK, and YES1, and AZD-5438 was identified as a novel RIPK1 inhibitor. By comparing cell-based and lysate-based data, the study distinguished direct target binding from downstream signaling effects.

Conclusion

This study demonstrates that proteome-wide thermal shift assays can characterize compound MoA at scale, identifying on-target engagement, off-target polypharmacology, and downstream pathway effects in a single experimental framework.

Thermal proteomics case study data showing PISA screen results for 96 compounds.

Figure 4 from Van Vranken et al. (2024): Cell- and lysate-based PISA as complementary approaches for assessing MoA. Reproduced under CC BY 4.0 license.

Bioinformatics Analysis That Turns Thermal Shifts into Mechanism Insights

Our bioinformatics pipeline is designed specifically for MoA interpretation, not just target list generation.

Tier 1 — Core Quantification and QC

Raw data processing, protein quantification, normalization, missing value handling, PCA and sample clustering. QC metrics include inter-replicate CV, temperature point correlation, and melting curve goodness-of-fit (R²).

Tier 2 — Thermal Shift Analysis

Melting curve fitting using four-parameter sigmoidal models, Tm calculation and shift determination, statistical testing (moderated t-test with FDR correction), and protein-level annotation.

Tier 3 — MoA Interpretation

This is where thermal data becomes mechanism insight: direct target classification (proteins shifted in both cell and lysate), downstream responder classification (proteins shifted in cell format only), polypharmacology network construction, pathway enrichment (GO, KEGG, Reactome), kinase enrichment analysis, and cross-method integration with phosphoproteomics, transcriptomics, or metabolomics data when available.

Software used: Spectronaut, DIA-NN, R/Bioconductor (limma, clusterProfiler, enrichR), custom Python pipelines for thermal curve fitting and MoA classification.

For combined approaches, see our LiP-MS + TPP and Phosphoproteomics activation mapping services.

Frequently Asked Questions

How does thermal proteomics distinguish direct target binding from downstream pathway effects?

By comparing thermal profiles from live-cell and lysate-based experiments. Direct binding events produce thermal shifts in both formats, while downstream effects only appear in live cells where the signaling context is intact. This comparison is built into our standard MoA workflow.

Can thermal proteomics work for membrane proteins or other difficult targets?

Yes. Thermal proteomics has been successfully applied to membrane proteins, including GPCRs and ion channels. For membrane protein targets, we optimize the lysis and solubilization conditions to preserve membrane protein thermal profiles. The PISA format is particularly well-suited for membrane proteins because it uses whole-cell lysates without fractionation.

How many compounds can be screened in a single thermal proteomics experiment?

Using the PISA format with TMTpro 16-plex or 18-plex labeling, we can profile up to 15 compounds (plus a control) in a single multiplexed experiment. For larger screens, we use a staggered design with common reference samples for cross-batch normalization.

How do thermal proteomics results compare with affinity-based methods like ABPP or pull-down?

Thermal proteomics and affinity-based methods are complementary. ABPP and pull-down detect direct physical binding but require compound modification and may miss low-affinity or transient interactions. Thermal proteomics detects both direct binding and downstream effects without compound modification, but does not directly measure binding affinity. We recommend using thermal proteomics for discovery and orthogonal methods for validation.

What validation do you recommend after thermal proteomics identifies candidate targets?

We recommend a tiered validation strategy: (1) thermal stabilization assay (PISA-based) or Western blot-based thermal shift for independent confirmation, (2) SPR or BLI for binding affinity measurement (if purified protein is available), (3) cellular functional assays (knockdown, overexpression, or mutant rescue) to confirm phenotypic relevance.

Can thermal proteomics data be integrated with other omics data?

Yes. Thermal proteomics integrates naturally with phosphoproteomics (to link target engagement to signaling changes), transcriptomics (to distinguish transcriptional from post-transcriptional effects), and metabolomics (to connect proteomic changes to metabolic phenotypes). We offer integrated multi-omics analysis packages for comprehensive MoA characterization.

References

Van Vranken JG, Li J, Mintseris J, et al. "Large-scale characterization of drug mechanism of action using proteome-wide thermal shift assays." eLife, 2024. DOI: 10.7554/eLife.95595 [CC BY 4.0]
Savitski MM, et al. "Tracking cancer drugs in living cells by thermal profiling of the proteome." Science, 2014. DOI: 10.1126/science.1255784
Mateus A, et al. "Thermal proteome profiling for interrogating protein interactions." Molecular Systems Biology, 2020. DOI: 10.15252/msb.20199232
Meissner F, et al. "The emerging role of mass spectrometry-based proteomics in drug discovery." Nature Reviews Drug Discovery, 2022. DOI: 10.1038/s41573-022-00409-3

Ready to uncover your compound's complete mechanism story with proteome-wide thermal profiling?

Contact our team to discuss your project and receive a detailed quotation.

Plan Your MoA Study View Sample Requirements

Disclaimer: All products and services provided by Creative Proteomics are for research use only (RUO). They are not intended for use in diagnostic, therapeutic, or clinical procedures.

Online Inquiry

Please submit a detailed description of your project. We will provide you with a customized project plan to meet your research requests. You can also send emails directly to for inquiries.