Nearly all pharmaceutical agents exert biological effects in a dose-dependent manner. In drug discovery and pharmacology, researchers have long struggled to connect observable cellular phenotypes with the intricate molecular mechanisms triggered by drugs. Core questions remain to be fully answered: which proteins directly bind to a candidate compound, what signaling pathways are activated or suppressed inside cells, and how cells undergo molecular and physiological reprogramming to reach drug-determined functional outcomes. Over the past decade, dramatic advances in the speed, sensitivity and accessibility of quantitative mass spectrometry have made large-scale, full dose-response experiments feasible across four critical research dimensions: target deconvolution, pathway engagement, proteome reprogramming and ultimate cellular consequences. Such approaches enable parallel acquisition of potency and effect size data for thousands of proteins and post-translational modification (PTM) sites, greatly boosting the depth of drug MoA research.

Traditional pharmacological assays often adopt a simple binary evaluation model and rely on arbitrarily selected high drug concentrations for proteomic or PTM profiling. This experimental design fails to distinguish multiple concurrent mechanisms of drugs with polypharmacology, masks differences in action intensity among regulated molecules, and severely limits the mechanistic interpretability of experimental data. In contrast, integrating the dose dimension into proteomic perturbation experiments brings three distinct strengths. It differentiates multiple drug targets and parallel pathways by potency variation, distinguishes subtle modulators from strong functional effectors via effect size comparison, and provides quantitative parameters for cross-analysis among compounds, pathways and cell states, far exceeding the limitations of simple "upregulation or downregulation" judgment.

The review first delves into the statistical hurdles of large-scale dose-response profiling. Mass spectrometry-based proteomic data are plagued by noisy quantification, missing values, non-sigmoidal response curves and heavy multiple testing pressure. The research team recommends applying a four-parameter log-logistic function to fit standard sigmoidal dose-response curves and introduces CurveCurator, a specialized software tool. Powered by multi-start regression and rigorous statistical thresholds, this tool classifies valid regulated curves, unregulated curves and ambiguous data, effectively controls the false discovery rate (FDR), and realizes reliable identification of true drug-responsive molecules from massive datasets.

Subsequently, the review systematically sorts out mature experimental techniques for decrypting drug MoA layer by layer, ranging from target binding to cellular phenotypes. For target deconvolution, it summarizes affinity enrichment, activity enrichment, stability-based detection and conformational accessibility analysis methods, including photoaffinity labeling (PAL), thermal proteome profiling (TPP), solvent-induced protein precipitation (iSPP) and limited proteolysis-coupled mass spectrometry (LiP-MS). Incorporating dose-response curves into these assays can filter non-specific binding, calculate apparent affinity and target occupancy, and accurately rank direct and indirect drug targets, overcoming the defects of single-dose experiments that tend to generate false positives.

In the research of pathway engagement, decryptM technology centered on PTM profiling (phosphorylation, acetylation, ubiquitinylation, etc.) stands out. By comparing the pEC₅₀ values of PTM sites, researchers can accurately divide different signaling pathways activated by drugs and clarify the transmission chain from target binding to downstream signal activation. Taking the multi-kinase inhibitor Dasatinib as an example, dose-resolved phosphoproteomics successfully revealed its divergent action characteristics in different cancer cell lines, which are determined by the dominant driver kinases of each cell. This methodology has also explained the synergistic mechanism of combined medication and the functional characteristics of various epigenetic drugs, showing broad application potential.

For proteome reprogramming and cellular adaptation, decryptE focuses on drug-induced changes in total protein abundance. Dose-response analysis can separate rapid direct effects of drugs on targets from slow adaptive reprogramming of cells, and distinguish primary MoA from secondary downstream reactions. Case studies of pomalidomide, methotrexate, carfilzomib and vorinostat prove that many protein expression changes are independent of transcriptional regulation. Notably, only about 25% of drugs cause abundance changes in their known target proteins, indicating that proteome alteration data are more suitable for analyzing cellular adaptation rather than direct target identification.

Linking multi-omics dose-response data to cellular phenotypes such as cell viability, apoptosis and cell cycle arrest is the key to translating molecular findings into clinical application. Time-resolved and dose-resolved joint analysis reveals a universal molecular cascade rule: molecular perturbations including target binding and pathway activation usually occur at lower drug concentrations and earlier time points than phenotypic changes. This rule helps researchers define the therapeutic window of drugs, identify dose ranges where effective target inhibition is achieved without extensive off-target interference, and evaluate drug safety and potential toxicity, providing important references for clinical dosing design and chemical safety assessment.

The review also puts forward practical guidelines for experimental design. The team suggests adopting a broad concentration range (no less than six orders of magnitude) with half-logarithmic spacing to fully capture complete sigmoidal response curves, and prioritizes increasing drug concentration points over repeated sample replicates. For data with missing values, targeted processing strategies are proposed based on different experimental scenarios. Meanwhile, the paper objectively analyzes the inherent limitations of current dose-resolved proteomics methods, including the dynamic range constraints of mass spectrometry, non-standard drug response curves, and interference from heterogeneous cell populations.

Looking ahead, the research team envisions multiple development directions for this field: combining dose-response measurements with subcellular fractionation, single-cell sequencing and protein turnover detection technologies; expanding experimental models from traditional two-dimensional cell cultures to spheroids, organoids and animal models. With mature conceptual frameworks, statistical tools and analytical technologies, system-wide dose-response proteomics will continue to drive innovation in mechanistic biology, drug discovery and translational pharmacology.

This insightful mini-review, titled Proteome-wide dose-response measurements for the characterization of drug mechanism of action, was published online on January 16, 2026, in Targetome.
doi: 10.48130/targetome-0025-0011