Biomarker to Target Translation by MS — Quantitative Proteomics, Thermal Profiling, and Multi-Omics Integration for Target Discovery and Validation

From correlative biomarker to validated drug target: a single MS-based pipeline bridging discovery proteomics, target deconvolution, and engagement confirmation.

The gap between biomarker discovery and target validation is one of the most resource-intensive transitions in drug discovery. A proteomic biomarker — a protein differentially abundant between disease and control — is a correlation, not a mechanism. Transforming that correlation into a druggable target requires: confirmation that the protein is functionally involved in disease biology, identification of compounds that engage it, quantification of target engagement at proteome scale, and prioritisation of the most therapeutically tractable candidates. Each step demands a different analytical modality, and the fragmentation of these capabilities across separate providers creates data compatibility gaps, sample handover losses, and project timeline delays. Our biomarker-to-target translation service integrates quantitative discovery proteomics, thermal proteome profiling (TPP and PISA), limited proteolysis-MS (LiP-MS), targeted MS assay development, and multi-omics computational integration within a single pipeline — enabling translational scientists to move from biomarker identification to validated, engagement-confirmed target candidates without switching platforms or providers.

Key Advantages:

End-to-end pipeline from biomarker discovery to target engagement — one platform, one project team, one data framework.
Proteome-wide target deconvolution by TPP, PISA, and LiP-MS — no target prefamiliarisation or antibody requirement.
Quantitative proteomics at cohort scale — DIA and TMT workflows for biomarker discovery across 100s of samples.
Target engagement quantification at the proteome level — measure compound-binding events across thousands of proteins simultaneously.
Multi-omics integration — proteomics, transcriptomics, and metabolomics data fused for target prioritisation with confidence scoring.
Orthogonal validation cascade — thermal profiling hits confirmed by native MS, HDX-MS, crosslinking MS, or SPR within the same project team.

Submit Your Target Discovery Project View FAQ

Biomarker to target translation by MS overview: a pipeline illustration showing the integrated workflow from quantitative proteomics biomarker discovery through thermal proteome profiling for target deconvolution, target engagement confirmation, and multi-omics target prioritisation — bridging the gap between correlative biomarkers and validated drug targets.

What Is Biomarker-Target Translation Service Modes Tech Comparison Sample Demo Case Study FAQ

What Is MS-Based Biomarker to Target Translation?

Biomarker-to-target translation is the systematic process of converting proteomic biomarker discoveries — differentially abundant proteins identified in disease versus control comparisons — into functionally validated, engagement-confirmed, and therapeutically prioritised drug targets. Mass spectrometry is the enabling technology at every stage of this pipeline because it provides the proteome-wide coverage, quantitative precision, and structural information needed to establish the causal chain from biomarker correlation to target validation.

The translation pipeline consists of four interconnected phases. In the discovery phase, quantitative proteomics (DIA or TMT) identifies proteins whose abundance changes with disease state, drug treatment, or phenotypic perturbation — producing a candidate biomarker list. In the deconvolution phase, thermal proteome profiling (TPP), proteome integral solubility alteration (PISA), or limited proteolysis-MS (LiP-MS) identifies which proteins are directly engaged by a compound of interest, distinguishing primary targets from downstream signalling effects. In the engagement phase, targeted MS assays (PRM/MRM) quantify target occupancy in dose-response and time-course experiments, confirming that compound binding at the target site produces the expected functional consequence. In the prioritisation phase, multi-omics data integration combines proteomics, transcriptomics, and metabolomics evidence with druggability assessment and disease association analysis to rank targets for entry into drug discovery programmes.

These capabilities are deployed as an integrated service within our Target → Drug Discovery platform, where biomarker-to-target translation provides the critical bridge between disease biology understanding and the initiation of screening campaigns for validated, engagement-confirmed targets.

Why MS for Biomarker-to-Target Translation

Discovery proteomics identifies the starting candidate list at proteome scale

Quantitative proteomics by DIA or TMT quantifies thousands of proteins across tens to hundreds of samples in a single experiment. The output — a ranked list of differentially abundant proteins with fold-change, statistical significance, and quantitative precision — is the starting point for the entire translation pipeline. No other technology provides this breadth of protein-level quantitation from complex biological samples without prefamiliarisation or antibody reagents.

Thermal profiling deconvolutes targets without target prefamiliarisation

TPP and PISA identify drug targets by measuring the change in protein thermal stability or solubility upon compound binding — a biophysical property that requires no prior knowledge of the target, no antibody, no labelled compound, and no immobilisation. The compound is incubated with a native cell lysate or intact cell, and the MS readout identifies all proteins whose thermal behaviour shifts upon compound binding. This unbiased, proteome-wide target discovery is the defining capability of the translation pipeline.

Target engagement quantification provides mechanistic confidence

A deconvoluted target is a hypothesis. Target engagement data — dose-dependent and time-dependent occupancy measured by quantitative MS — converts that hypothesis into a confirmed mechanism. PRM and MRM assays targeting the specific protein or proteotypic peptides quantify target occupancy in compound-treated samples, establishing the relationship between compound concentration, target binding, and phenotypic response. This three-way linkage is the gold standard for translational target validation.

LiP-MS reveals compound binding sites at proteome scale

Limited proteolysis-MS (LiP-MS) identifies not only which protein a compound binds to, but which region of the protein is bound. Compound-induced conformational changes alter the proteolysis pattern of the target protein, and the differential peptide fragments identify the binding footprint at amino-acid resolution. For targets without known structures or for allosteric binding sites, this provides structural information that guides medicinal chemistry before a co-crystal structure is available.

Multi-omics integration prioritises targets by confidence score

A protein identified by thermal profiling may be a true target, a downstream effector, or a bystander with coincidental thermal shift. Multi-omics integration resolves this ambiguity by correlating proteomic target hits with transcriptomic evidence (is the target expressed in the disease-relevant tissue?), genetic evidence (are loss-of-function variants associated with the disease phenotype?), and metabolomic evidence (does target engagement produce the expected metabolite changes?). Targets with concordant evidence across multiple omics layers are prioritised for entry into drug discovery.

Orthogonal validation cascade confirms targets before programme entry

Every target identified through thermal profiling or LiP-MS is passed through an orthogonal validation cascade: target-appropriate biochemical or cellular assay confirmation, native MS for binding stoichiometry and affinity where the target can be purified, and SPR or HDX-MS for kinetic and structural characterisation of priority targets. This cascade ensures that only targets with confirmed compound engagement — at the right site, with the right affinity, in the right biological context — advance into hit-to-lead programmes.

Service Modes — MS-Based Biomarker-to-Target Translation

We offer four service modes that correspond to the four phases of the biomarker-to-target translation pipeline. Modes can be deployed as standalone services for projects at a specific stage, or combined into an end-to-end campaign that spans from discovery proteomics through target validation and engagement confirmation.

MODE 1

Biomarker Discovery by Quantitative Proteomics

Cohort-scale quantitative proteomics by DIA or TMT to identify differentially abundant proteins associated with disease state, drug treatment, or phenotypic perturbation. Samples — plasma, serum, CSF, tissue lysates, or cell culture — are processed through a standardised proteomics workflow with stringent quality control at every step. The output is a statistically ranked list of candidate biomarkers with fold-change, p-value, and false discovery rate.

Format: DIA (data-independent acquisition) for label-free quantification across unlimited samples; TMTpro (16- or 18-plex) for multiplexed quantification with internal reference channel.
Scale: 10–500+ samples per project; typical depth of 4,000–8,000 protein groups per sample (DIA) or 6,000–9,000 per multiplex set (TMT).
Output: Ranked biomarker list with statistical parameters; PCA and clustering visualisations; enrichment analysis (GO, KEGG, Reactome); candidate target shortlist for downstream validation. For deeper proteome coverage in specific biological contexts, our proteomics drug-response profiling service provides extended fractionation workflows.

MODE 2

Drug Target Deconvolution by Thermal Profiling and LiP-MS

Unbiased, proteome-wide identification of compound targets using thermal proteome profiling (TPP), proteome integral solubility alteration (PISA), and limited proteolysis-MS (LiP-MS). The compound of interest — a phenotypic hit, natural product, covalent probe, or clinical candidate — is incubated with cell lysate or intact cells, and the proteome-wide response to compound binding is measured by quantitative MS. Proteins that undergo a significant change in thermal stability (TPP/PISA) or proteolysis pattern (LiP-MS) upon compound treatment are identified as candidate targets.

Format: TPP (10-temperature melting curve, DIA or TMT readout); PISA (single-temperature solubility shift, label-free DIA readout); LiP-MS (proteolysis pattern comparison with and without compound).
Sample types: Cell lysates, intact cells, tissue homogenates; compatible with any compound that can be delivered to the native proteome.
Output: Ranked target list with thermal shift significance, melting temperature (Tm) values, and dose-response thermal shift data; LiP-MS binding site peptide maps where applicable. For dedicated TPP-LiP integrated workflows, see our LiP-MS + TPP service. For high-throughput thermal profiling with DIA readout, our DIA + TPP platform provides faster acquisition and deeper coverage.

MODE 3

Target Engagement and Occupancy Quantification

Quantitative targeted MS assays (PRM or MRM) for measuring target engagement in compound-treated biological systems. Once a target is identified by thermal profiling, a targeted MS assay is developed for that specific protein using proteotypic peptides. The assay quantifies target abundance and, where applicable, post-translational modification state (e.g. phosphorylation at the active site) across compound concentration ranges and time points. Target occupancy — the fraction of target protein bound by compound — is inferred from the dose-response relationship between compound concentration and target modification or stabilisation.

Format: PRM (parallel reaction monitoring) on high-resolution Orbitrap or Q-TOF; MRM (multiple reaction monitoring) on triple-quadrupole MS for higher-throughput quantification.
Quantification: Stable isotope-labelled peptide standards (SIS) for absolute quantification where required; label-free relative quantification for screening-scale studies.
Output: Target engagement dose-response curve with EC₅₀ or IC₅₀; time-course occupancy data; comparison of target engagement with phenotypic response for mechanism-of-action confirmation.

MODE 4

Multi-Omics Target Prioritisation and Integration

Computational integration of proteomic target discovery data with transcriptomic, genetic, metabolomic, and structural evidence to rank and prioritise targets for entry into drug discovery programmes. The integration platform combines: thermal profiling significance scores, disease association evidence (from GWAS, phenome-wide association studies, and rare-variant burden tests), tissue expression specificity, druggability assessment (binding pocket prediction, lipophilicity, target class family), and structural feasibility (crystallised structure availability, predicted AlphaFold confidence).

Format: Custom bioinformatics pipeline integrating multi-omics datasets; target confidence matrix with weighted evidence scores across each data type.
Input: Proteomics data from Mode 1 or Mode 2; optional transcriptomics (RNA-seq), genomics (GWAS summary statistics), metabolomics, and literature-curated data.
Output: Ranked target list with multi-omics confidence scores; target-disease association evidence summaries; druggability assessment for each prioritised target; visualisation of multi-omics evidence matrix. For detailed multi-omics integration using AI-based approaches, our deep multi-omics integration platform provides advanced machine learning models for target discovery. For interaction network context, see our interactomics service.

Analytical Workflow

Five stages from sample receipt to prioritised, validated target list:

Project scoping and experimental design

The project team reviews the customer's biomarker or compound data, defines the translation path (biomarker → target, compound → target, or both), and designs the experimental cascade. For biomarker-initiated projects: sample cohort size, disease versus control comparison design, proteomics acquisition strategy (DIA vs TMT), and statistical power analysis. For compound-initiated projects: compound properties (solubility, cell permeability, incubation conditions), thermal profiling format (TPP vs PISA vs LiP-MS), and required number of biological replicates.

Discovery proteomics or thermal profiling acquisition

For biomarker discovery mode: LC-MS/MS analysis of each sample by DIA or TMT with standardised quality control (internal standards, pooled reference samples, system suitability monitoring). For thermal profiling mode: compound-treated and vehicle-treated samples are fractionated across the temperature gradient (TPP) or solubility separation (PISA), and each fraction is analysed by quantitative DIA or TMT MS. Typical acquisition time: 2–5 days per 50-sample cohort (DIA) or per 10-temperature TPP experiment.

Target identification and statistical analysis

Raw MS data are processed through the quantification pipeline (DIA-NN, Spectronaut, or MaxQuant for DIA; Proteome Discoverer or MSFragger for TMT). Differentially abundant proteins (biomarker mode) are identified by moderated t-test or ANOVA with FDR correction. Thermal shift targets (TPP mode) are identified by melting curve fitting (TPP R package or NPARC) with significance thresholds; PISA targets are identified by solubility ratio comparison. LiP-MS targets are identified by differential peptide abundance analysis between compound-treated and untreated samples.

Target engagement confirmation and orthogonal validation

Priority targets from the discovery phase enter the engagement confirmation workflow. PRM or MRM assays are developed for the target proteins, and dose-response and time-course experiments are performed to quantify target engagement. Orthogonal validation by native MS, SPR, or HDX-MS is initiated for the highest-priority targets. A target confidence matrix is built, incorporating: thermal shift significance, dose-response EC₅₀ concordance, orthogonal binding confirmation, and cellular phenotypic correlation.

Multi-omics integration and target prioritisation report

All data streams — discovery proteomics, thermal profiling, target engagement, orthogonal validation — are integrated into a final target prioritisation report. Multi-omics evidence from public databases (Open Targets, GWAS Catalog, GTEx, DepMap) is overlaid to assess disease association, tissue specificity, genetic validation, and druggability. Each target receives a confidence score across five evidence dimensions. The final report includes a ranked target list with full evidence traceability, recommended follow-up strategy, and estimated resource requirements for advancing each target into hit identification.

Biomarker to target translation analytical workflow: project scoping and experimental design, discovery proteomics or thermal profiling MS acquisition, target identification and statistical analysis, target engagement confirmation by PRM/MRM with orthogonal validation, and multi-omics integration for target prioritisation report with confidence scoring and druggability assessment.

Applications by Drug Discovery Stage

The biomarker-to-target translation pipeline addresses distinct needs at each stage of the drug discovery lifecycle, from early target discovery through clinical candidate characterisation.

Phenotypic hit deconvolution and target identification

When a phenotypic screen produces a hit compound with a desired cellular phenotype but an unknown mechanism, thermal profiling (TPP/PISA/LiP-MS) identifies the protein target(s) responsible. The compound is incubated with the relevant cell type or lysate, and the proteome-wide thermal shift or proteolysis pattern is measured. Targets are ranked by significance and dose-response consistency. This approach is applicable to any compound class — small molecules, natural products, covalent probes, and PROTACs — and requires no prior target hypothesis.

Output: Ranked target list with thermal shift significance; LiP-MS binding site information where detected; dose-response target engagement confirmation for priority candidates.

Biomarker-to-target transition for translational programmes

Proteomic biomarker studies frequently identify dozens to hundreds of differentially abundant proteins. Prioritising which of these proteins are causal drivers (drug targets) versus downstream consequences (biomarkers) is the central challenge of translational proteomics. The integration of biomarker discovery with thermal profiling and multi-omics evidence enables systematic prioritisation: proteins that are both differentially abundant in disease and thermally stabilised by a reference compound receive the highest confidence scores.

Output: Integrated biomarker-target prioritisation matrix; causal versus correlative classification for each candidate; compound-target engagement data for druggable candidates.

Target engagement and selectivity profiling in lead optimisation

During lead optimisation, confirming that a compound series engages its intended target in the disease-relevant cellular context — and quantifying the selectivity over off-targets — is essential for compound advancement. Thermal profiling provides proteome-wide selectivity data in a single experiment: the compound is incubated with the target cell type, and all proteins that show a thermal shift are identified. Off-targets are ranked by affinity relative to the primary target, enabling medicinal chemistry teams to address selectivity liabilities early.

Output: Proteome-wide selectivity profile with target and off-target ranking; dose-response target engagement curves in the disease-relevant cell type; correlation of target occupancy with phenotypic response.

Natural product and covalent probe target discovery

Natural products with known phenotypic activity but unknown targets — and covalent probes designed to engage specific protein classes — are both optimally deconvoluted by thermal profiling or LiP-MS. For natural products, the unbiased nature of thermal profiling is critical: the actual target may be unrelated to the presumed mechanism. For covalent probes, LiP-MS can identify the modified cysteine residue or the conformational change induced by covalent binding, providing structural information that guides probe optimisation.

Output: Deconvoluted target list with thermal shift or LiP-MS significance; covalent adduct identification by intact protein MS or target peptide MS/MS; structural insight into binding mode where LiP-MS peptide-level data are obtained.

PROTAC and molecular glue target engagement confirmation

PROTACs and molecular glues induce targeted protein degradation through ternary complex formation — a mechanism that cannot be confirmed by standard biochemical assays. Thermal profiling detects the stabilisation or destabilisation of both the target protein and the E3 ligase upon PROTAC treatment, providing a proteome-wide readout of ternary complex formation and degradation selectivity. The thermal shift of the target-E3 ligase interaction surface can be distinguished from off-target thermal shifts by the dose-response concordance between the two proteins.

Output: Target and E3 ligase thermal shift confirmation; degradation selectivity profile across the proteome; dose-response correlation between ternary complex formation and target degradation. For PROTAC-specific ternary complex profiling, see our dedicated services for ubiquitinomics by MS and proximity labelling.

Mechanism-of-action elucidation for clinical-stage compounds

For compounds that advance into clinical development with an incompletely understood mechanism — or for which unexpected clinical efficacy or toxicity is observed — thermal profiling and LiP-MS provide retrospective mechanism-of-action data. Archived cell lines, patient-derived samples, or surrogate tissues can be treated with the compound and analysed for proteome-wide targets. The resulting data can explain efficacy (identification of the true therapeutic target), toxicity (off-target identification), or resistance mechanisms (pathway adaptation signatures).

Output: Clinical-stage compound target deconvolution data; off-target toxicity mechanism hypothesis; resistance pathway identification; recommendation for combination therapy or indication expansion strategy.

Technology Comparison: MS-Based Target ID vs Alternative Approaches

Platform	Detection Principle	Unbiased (Proteome-Wide)?	Target Preknowledge Required?	Binding Site Info?	Applicable Compound Classes	Cellular Context
TPP / PISA / LiP-MS (this service)	Thermal stability or proteolysis shift upon compound binding	✅ Yes	None	✅ Yes (LiP-MS)	All soluble compounds, covalent probes, natural products, PROTACs	Lysate or intact cell
Affinity Purification-MS (AP-MS)	Pull-down of tagged compound or bait protein	⚠️ Limited — requires immobilised compound or antibody	Compound modification or tag required	⚠️ Partial — identifies bound protein, not binding site	Compounds amenable to immobilisation or biotinylation	Lysate only
CRISPR / Genetic Screens	Gene knockout or activation phenotype correlation	✅ Yes (genome-wide)	None	❌ No — gene-level only	Compounds with measurable fitness or reporter phenotype	Intact cell
Computational Target Prediction	Chemical similarity, structural docking, or ML-based prediction	✅ Yes (in silico)	Compound structure required	⚠️ Partial — predicted binding mode	All compounds with known structure	In silico only — no experimental confirmation
Cellular Thermal Shift Assay (CETSA) — Single Target	Immunodetection of target melting after compound treatment	❌ No — antibody required per target	Target antibody required; limited to known targets	❌ No	Compounds with known or suspected target	Intact cell or lysate
Photoaffinity Labelling (PAL)	UV-crosslinking of photoreactive compound analogue	⚠️ Partial — requires photoreactive compound analogue	Compound must be synthetically modified	✅ Yes — crosslinked peptide identified by MS	Compounds amenable to photoreactive analogue synthesis	Lysate or intact cell

Sample Requirements

Component	Format Options	Recommended Input	Minimum Input	Key Notes
Cell Lysate (TPP/PISA/LiP-MS)	Native cell lysate in PBS or MS-compatible buffer	2–5 mg total protein per condition	1 mg per 10-temperature TPP experiment	Provide cell type, lysis method, and protease/phosphatase inhibitor status; avoid MS-incompatible detergents
Intact Cells (cellular TPP)	Live cell suspension or adherent culture	5 × 10⁶ cells per condition	1 × 10⁶ cells	Compound pre-incubation at 37°C for 1 hr; provide vehicle control (DMSO ≤0.1%)
Biofluid (biomarker discovery)	Plasma, serum, CSF, or other biofluid	100 µL per sample (plasma/serum); 500 µL (CSF)	20 µL (plasma); 100 µL (CSF)	Collect in standard clinical tubes; avoid haemolysis; aliquot and freeze at −80°C within 2 hours
Tissue (biomarker discovery)	Snap-frozen or cryopreserved tissue	10–50 mg per sample	5 mg	Provide disease and control tissue matched by age, sex, and tissue collection site; avoid OCT compound if possible
Test Compound (target deconvolution)	Solution in DMSO or aqueous buffer	10–50 µL at 10 mM	5 µL at 1 mM	Provide known solubility limit; DMSO stocks preferred; if compound is light-sensitive or oxidation-prone, provide under inert atmosphere
Control Compound (optional)	Solution in DMSO or aqueous buffer	10 µL at 10 mM	5 µL at 1 mM	Known target-engaging compound for method validation; positive control for thermal profiling assay performance assessment

For biomarker discovery projects involving clinical samples, we require documentation of ethical approval and informed consent. Sample shipment should be on dry ice with temperature logging. For projects combining multiple omics modalities (proteomics + transcriptomics + metabolomics), we recommend splitting a single sample aliquot at the time of collection rather than using separate collections, to minimise pre-analytical variability.

Deliverables

Biomarker discovery report: differentially abundant protein list with fold-change, p-value, FDR; PCA and clustering analysis; GO/KEGG/Reactome enrichment; candidate target shortlist with disease association evidence.
Thermal profiling target deconvolution report: ranked target list with melting temperature (Tm) values, thermal shift significance (TPP mode); solubility ratio comparison (PISA mode); dose-response thermal shift data for confirmed targets.
LiP-MS binding site report: differential peptide abundance maps; binding footprint identification at peptide resolution; conformational change visualisation on protein structure (where available).
Target engagement quantification report: PRM/MRM dose-response curves with EC₅₀ or IC₅₀; time-course occupancy data; comparison of target engagement with phenotypic response.
Multi-omics target prioritisation matrix: ranked target list with confidence scores across five evidence dimensions (thermal profiling, disease association, genetic validation, druggability, structural feasibility).
Orthogonal validation data: native MS binding confirmation (Kd, stoichiometry); SPR sensograms and affinity data; HDX-MS binding epitope maps (where applicable).
Raw MS data files in standard formats (mzXML, .raw) and processed quantification tables.
Written translation strategy report: integrated interpretation of all data streams, recommended target advancement path, estimated timeline and resources for hit identification campaign initiation.

Representative Results

Volcano plot from quantitative proteomics biomarker discovery: log2 fold-change on x-axis versus -log10 p-value on y-axis, with significantly upregulated proteins in red and downregulated in blue, labelled with candidate target protein names.

Biomarker discovery by DIA quantitative proteomics

Volcano plot from a DIA-based biomarker discovery study comparing disease (n=50) versus healthy control (n=50) plasma samples. Over 4,500 protein groups were quantified across the cohort. Proteins with fold-change >1.5 and p-value<0.05 (FDR-corrected) are highlighted — 127 upregulated (red) and 89 downregulated (blue) in the disease group. Candidate target proteins are labelled, and the top 20 candidates by combined fold-change and significance are annotated with gene names. This ranked list enters the target prioritisation pipeline for thermal profiling and multi-omics integration.

$Thermal proteome profiling (TPP) melting curves: protein melting curves (fraction non-denatured versus temperature) for control (vehicle, blue) and compound-treated (red) samples, showing thermal shift for identified target proteins with Tm shift annotation.$

Target deconvolution by TPP: melting curve shifts confirm target engagement

Representative thermal proteome profiling data from a compound target deconvolution experiment. Melting curves (fraction of non-denatured protein versus temperature) are shown for three identified target proteins. Control curves (vehicle treatment, blue) and compound-treated curves (red) are overlaid. A significant thermal shift (ΔTm > 2°C, p < 0.01) is observed for each target. The magnitude and direction of the shift — stabilisation (rightward shift) for two targets and destabilisation (leftward shift) for one — are consistent with direct binding and compound-induced conformational change, respectively. The full TPP experiment quantified 6,200 proteins and identified 14 significant thermal shifts, of which 5 were confirmed by dose-response TPP and 3 by orthogonal SPR.

Multi-omics target prioritisation matrix: a heatmap-style visualisation showing confidence scores across five evidence dimensions (thermal profiling significance, disease association, genetic validation, druggability, structural feasibility) for the top 10 candidate targets, with overall prioritisation score.

Multi-omics target prioritisation: evidence integration and confidence scoring

Target prioritisation matrix from a multi-omics integration pipeline. The top 10 candidate targets are shown as rows, with five evidence dimensions as columns: thermal profiling significance (ΔTm and p-value from TPP), disease association (GWAS and phenome-wide association z-score), genetic validation (DepMap dependency score and rare-variant burden), druggability (predicted pocket quality and target class tractability score), and structural feasibility (crystallised structure or AlphaFold confidence). Each cell is colour-coded from green (strong evidence) to red (no evidence). The overall prioritisation score (rightmost column) integrates the five dimensions into a single rank. Targets with concordant evidence across multiple dimensions — thermal shift + genetic dependency + high druggability — are recommended for immediate entry into hit identification campaigns.

Case Study: Fully Automated Thermal Proteome Profiling Identifies Kinase Inhibitor Targets and Off-Targets at Proteome Scale

Wu Q., Zheng J., Sui X., et al. "High-throughput drug target discovery using a fully automated proteomics sample preparation platform." Chemical Science. 2024;15:2833–2847. https://doi.org/10.1039/d3sc05937e Open Access (CC BY-NC 3.0).

Background

Thermal proteome profiling (TPP) is a powerful method for unbiased drug target identification, but its adoption in drug discovery has been limited by the complexity and low throughput of conventional TPP workflows — which typically involve manual sample preparation across 10 temperature points per compound, multiple liquid handling steps, and extensive MS acquisition time. The authors sought to develop a fully automated, high-throughput TPP platform capable of processing multiple compounds in parallel for proteome-wide target and off-target identification at a scale practical for drug discovery project support.

Methods

The autoSISPROT platform integrates a 96-channel automated sample preparation system with single-temperature TPP (STPP) and DIA-based quantitative proteomics. Instead of the classical 10-temperature melting curve format, STPP uses a single temperature near the median melting temperature of the proteome — compounds that bind and stabilise a target shift its melting curve and alter the soluble protein fraction at that temperature. The autoSISPROT system processes 96 samples in parallel, performing reduction, alkylation, and digestion in under 3 hours with minimal manual intervention. The authors applied the platform to profile 20 kinase inhibitors — including clinical and preclinical compounds — in K562 leukaemia cells, identifying both intended targets and off-targets for each compound.

Results

The autoSISPROT-STPP platform quantified over 5,000 proteins per experiment with a median CV of less than 10%, demonstrating reproducibility comparable to classical TPP workflows. Across 20 kinase inhibitors, the platform correctly identified the primary kinase targets (e.g., BCR-ABL for dasatinib, EGFR for gefitinib, MEK1/2 for trametinib) and discovered known and novel off-targets — including off-target kinases and non-kinase proteins. The single-temperature format achieved greater than 10-fold improvement in throughput over classical TPP, enabling a 20-compound profiling campaign to be completed in a fraction of the time required by conventional methods. Target engagement was confirmed by dose-response STPP for selected compound-target pairs, and orthogonal validation by immunoblotting and kinase activity assays confirmed the top deconvoluted targets.

Significance for MS-Based Target Discovery

This study establishes that thermal proteome profiling can be deployed at a throughput and reproducibility level suitable for routine drug discovery project support — not just as a specialised deconvolution method for疑难 cases, but as a standard target identification and selectivity profiling tool applicable throughout hit-to-lead and lead optimisation. The integration of automation, single-temperature format, and DIA quantification provides a practical template for making proteome-wide target deconvolution accessible to drug discovery teams who need target and off-target data on project-relevant timelines.

Figure from Wu et al. 2024, Chemical Science, showing the autoSISPROT-STPP thermal proteome profiling workflow and target identification results for 20 kinase inhibitors, including melting curve shifts and target/off-target ranking.

Figure 2 from Wu et al. 2024 (Chemical Science, DOI: 10.1039/d3sc05937e, CC BY-NC 3.0). autoSISPROT-STPP workflow schematic and representative thermal profiling results for kinase inhibitor target deconvolution.

FAQ

Frequently Asked Questions

Q: What types of compounds are compatible with TPP and LiP-MS for target deconvolution?

Any compound that binds to a protein in a cellular context and alters its thermal stability or proteolysis pattern is compatible. This includes reversible small-molecule inhibitors, covalent inhibitors, natural products, PROTACs, molecular glues, stabilised peptides, and macrocycles. Compounds must be soluble in the assay buffer at the screening concentration (typically 1–50 µM), and cell-permeable compounds are required for intact-cell TPP experiments. Highly insoluble compounds, very large biologics (>30 kDa), and compounds requiring metabolic activation before target engagement may require alternative formats (lysate TPP or specific metabolic pre-incubation). We assess compound suitability during the feasibility phase and recommend the appropriate thermal profiling format (TPP, PISA, or LiP-MS) based on compound properties.

Q: How many targets can be identified from a single thermal profiling experiment?

A single TPP or PISA experiment typically quantifies 4,000–7,000 proteins, of which 5–50 may show significant thermal shifts upon compound treatment — depending on compound selectivity, binding affinity, incubation conditions, and the stringency of the statistical threshold. Highly selective compounds targeting abundant proteins may produce 1–3 significant thermal shifts. Multi-target drugs, promiscuous compounds, and compounds targeting protein families with conserved binding sites may produce 20–50 or more. Each significant thermal shift is a candidate target that progresses to dose-response confirmation and orthogonal validation. False positives from compound aggregation, assay interference, or off-target thermal effects are controlled by dose-response TTP and buffer-dependency checks.

Q: Can thermal profiling distinguish between direct targets and downstream effectors?

Yes, with appropriate experimental design. Proteins whose thermal shift depends on compound binding — stabilised or destabilised through direct interaction with the compound — exhibit a dose-dependent thermal shift that saturates at compound concentrations consistent with the binding affinity. Downstream effectors may show a thermal shift that is not dose-responsive (if the shift results from a signalling cascade rather than direct binding) or that occurs only at compound concentrations above the target saturation point. In intact-cell TPP, we recommend time-course experiments: a direct target shift should appear within minutes of compound addition, while downstream effectors may require longer incubation. LiP-MS provides an additional layer of discrimination by identifying the specific peptide region protected from proteolysis by compound binding — direct evidence of a physical interaction site.

Q: How does PISA differ from classical TPP, and which should I choose?

Classical TPP measures protein melting curves across 10 temperature points and calculates the melting temperature (Tm) shift for each protein. PISA (Proteome Integral Solubility Alteration) measures the soluble protein fraction at a single temperature and calculates a solubility ratio between compound-treated and control samples. PISA offers higher throughput (fewer MS runs per experiment) and simpler data analysis (ratio comparison versus curve fitting) but provides less information — it does not yield Tm values or curve shape parameters that can help classify binding mode. Classical TPP is recommended for: detailed target characterisation, distinguishing stabilisation from destabilisation, and targets where the thermal shift is subtle. PISA is recommended for: high-throughput multi-compound profiling campaigns, early-stage target identification where speed is prioritised over mechanistic detail, and samples where protein quantity is limited. Both formats are available within our service.

Q: What statistical power is needed for a biomarker discovery proteomics study?

For DIA-based quantitative proteomics, the minimum recommended cohort size is 10 samples per group (disease and control) for detecting proteins with a fold-change ≥2 at 80% power and α = 0.05 (FDR-corrected). For detecting smaller fold-changes (≥1.5), we recommend a minimum of 25–30 samples per group. For TMT-based studies, the multiplexing format (16- or 18-plex) limits per-group sample numbers within each plex, but the inclusion of a common reference channel across multiple plexes enables larger cohort comparisons. We provide statistical power analysis during the project scoping phase and recommend cohort sizes matched to the expected effect size distribution for the specific sample type and disease indication. Pilot studies using 5–10 samples per group are recommended for novel indications where the expected effect size is unknown.

Q: How long does a complete biomarker-to-target translation project take?

Timelines depend on project scope and the number of modes deployed. A single-mode project — biomarker discovery only or thermal profiling for a single compound — is typically completed in 4–8 weeks from sample receipt to final report. An integrated two-mode project (biomarker discovery + thermal profiling for prioritised targets) requires 8–14 weeks. A full end-to-end project spanning all four modes (biomarker discovery, target deconvolution, target engagement quantification, and multi-omics prioritisation) for a single disease indication and 3–5 prioritised compounds is typically completed in 12–20 weeks. We provide a detailed project timeline with milestones during the project scoping phase. Expedited timelines are available for time-critical programme decisions, typically compressing the timeline by 25–40% by running discovery and deconvolution streams in parallel.

References

Wu Q., Zheng J., Sui X., et al. High-throughput drug target discovery using a fully automated proteomics sample preparation platform. Chem Sci. 2024;15:2833–2847.
Yang L., Guo C.W., Luo Q.M., et al. Thermostability-assisted limited proteolysis-coupled mass spectrometry for capturing drug target proteins and sites. Anal Chim Acta. 2024;1312:342755.
Ser Z., Sobota R.M. Proteome integral solubility alteration via label-free DIA approach (PISA-DIA). Proteomics. 2025;25(3):e2400147.
George A.L., Sidgwick F.R., Watt J.E., et al. Comparison of quantitative mass spectrometric methods for drug target identification by thermal proteome profiling. J Proteome Res. 2023;22(8):2629–2640.

Design Your Biomarker-to-Target Translation Project with the MassTarget Team

Tell us your project stage — whether you have a biomarker panel needing target validation, a compound needing target deconvolution, or a multi-omics dataset needing integrated prioritisation — and our scientists will design a translation strategy matched to your programme timeline and decision needs.

Start your target discovery inquiry

For Research Use Only (RUO). Not intended for diagnostic, therapeutic, or clinical decision-making purposes. Creative Proteomics services are designed to support preclinical research, drug discovery, and mechanism of action studies only.

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.