High resolution ES-DMA quantifies AAV capsid DNA content by electrical mobility to mass correlation Authors: Preston Dennett, Lohra M Young, Benjamin E Draper, Jerome Schmitt, Martin Jarrold, Luis Javier Perez-Lorenzo Published: 2026 ABSTRACT AbstractRecombinant Adeno-Associated Virus (rAAV) is the leading viral vector platform for gene therapy. A persistent challenge in rAAV manufacturing and quality control is accurate assessment of the DNA content of purified viral capsids. Empty particles, as well as ‘partial’ and over-filled species commonly contaminate even highly purified rAAV preparations. These impurities contribute to high production costs and reduced therapeutic efficacy, resulting in higher required doses and increased risks to patient health. Here we present high-resolution electrospray differential mobility analysis (ES-DMA) as a novel benchtop analytical modality for rAAV vector characterization. Sampling microliter volumes of analyte, this method achieves angstrom-scale particle sizing which we demonstrate is sufficient to resolve DNA encapsidation by rAAVs via differential electrical mobility. We benchmark this approach against charge detection mass spectrometry (CD-MS), an emerging gold standard for rAAV analytics, and identify a robust near-linear correlation between electrical mobility and mass of aerosolized rAAV particles. This relationship enables rapid identification and relative quantification of empty, full, and partial / over-filled rAAV capsids by ES-DMA. Subject terms: Biophysical methods, Genetic vectors Introduction Adeno-associated virus (AAV) is a ~ 25 nm non-enveloped icosahedral parvovirus encapsidating a 4.7 kilobase (kb) ssDNA genome [1, 2]. Due to its non-pathogenic nature, relatively low immunogenicity, and broad tissue tropism, AAV has become established as the most common viral vector employed in gene therapy [3]. With 7 AAV-based gene therapy products currently approved by the FDA for treatment of various rare genetic diseases, the AAV platform appears well-positioned to expand to broader indications and larger numbers of patients [4]. However, despite the promise of potentially curative treatments and extensive effort, wider public health impact of AAV-based gene therapies has yet to be realized, partly due to low therapeutic efficacy and the extremely high costs of developing and manufacturing these complex biologic drug products [5] While AAV vectors are generally considered to have a strong safety profile, several patient deaths have occurred, particularly when administered at high doses [6]. Heterogeneity in rAAV capsid DNA content has been identified as a potential key contributor to these safety risks by undermining reproducibility and interpretation of both pre-clinical studies and clinical trial data [7, 8]. Indeed, reliable determination and optimization of rAAV capsid “empty/full ratios” has long remained a persistent manufacturing challenge, with relatively large quantities of empty capsids frequently co-purifying with the desired ‘full’ particles. These empty capsids contribute to anti-AAV immunogenicity and associated toxicity in patients, likely without providing any therapeutic benefit [9, 10]. Despite their contribution to overall viral load, empty capsids are not typically included in dosage calculations which rely on estimates of DNA content and similarly, the abundance of empty capsids is generally not reported in clinical trial data [7, 8]. Beyond empty/full rAAVs, recent work has recognized the urgent need for novel analytical technologies capable of detecting and quantifying “partial” and “over-filled” capsids as well [11, 12]. These contaminant particles can contain poorly characterized and potentially hazardous genetic material, including truncated therapeutic transgenes and/or plasmid DNA vector components [13, 14]. Notably, because they often encapsidate some or all of the desired DNA transgene, these dysfunctional species can also contribute to distorted estimates of AAV titer and empty/full particle ratios, which are often determined by quantitative PCR (qPCR) and used for calculating patient doses [8]. If not reliably identified and removed, these species thereby artificially inflate the proportion of “full” capsids while undermining product potency and safety. Additionally, heterogenous contaminants may also arise from non-genomic cargo, including host-cell proteins (HCPs), further complicating characterization of AAV preparations and underscoring the need for orthogonal analytical methods capable of rapidly profiling particle populations [12]. Several technologies have been developed and applied to attempt to address these issues in AAV manufacturing and quality control [15]. Negative stain electron microscopy (EM) is a long-established method for assessing empty/full particle ratios of purified AAV samples but is low throughput and unable to reliably quantify partial and over-filled species [16]. Size exclusion chromatography combined with multi-angle light scattering (SEC-MALS) is able to provide information on several critical quality attributes of AAVs but estimates of empty/full ratios are similarly distorted by the presence of unresolved partial / over-filled capsids [16]. Recently, mass photometry was shown to be capable of detecting partial / over-filled capsids, however quantification remains sensitive to sample formulation and impurities [17, 18]. Analytical ultracentrifugation (AUC) is likely the highest resolution method currently available for detection of partial / over-filled species but requires relatively large sample volumes and long run times using specialized equipment and complex analysis [19, 20]. Additionally, charge detection mass spectrometry (CD-MS) represents an emerging gold standard for AAV analytics, providing unambiguous mass analysis of high molecular weight protein assemblies, however these systems have only recently become commercially available [21–23]. Despite this substantial progress, application of multiple orthogonal technologies in parallel will likely continue to be essential for comprehensive characterization of rAAV vectors [8, 24]. Here we sought to determine whether gas-phase electrical mobility, a proxy for particle size, quantified via electrospray differential mobility analysis (ES-DMA), could be applied to assess AAV capsid content. Specifically, we hypothesized that upon electrospray ionization and aerosolization in ES-DMA analysis, AAV particles would undergo partial dehydration and compaction around the encapsidated DNA cargo, as reported previously for several RNA viruses [25–27]. Given a sufficiently high resolution sizing tool, we anticipated that this process would yield measurable differences in rAAV capsid electrical mobility corresponding to the underlying DNA content. Electrospray differential mobility analysis (ES-DMA) is a well-established method for nanoparticle characterization, enabling sub-nanometer sizing and quantification [28, 29]. In this technique, analytes are transferred from an aqueous electrolyte solution into the gas phase via nano-electrospray ionization. The resulting charged droplets undergo solvent evaporation and charge equilibration in a neutralizing atmosphere [29]. Ultimately a substantial fraction of the analyte features a single charge, thereby reducing spectral complexity [30, 31]. Following electrospray ionization and charge reduction, the predominately singly charged particle stream enters the differential mobility analyzer (DMA), where they are separated according to their electrophoretic mobility under the combined influence of a flow field (“sheath flow”) and a perpendicular electric field (“DMA voltage”) [32]. As the inner electrode voltage is varied, the DMA operates as a tunable mobility band-pass filter, transmitting only particles of a given electrical mobility to the detector. A condensation particle counter (CPC) connected to the DMA outlet provides single particle quantification by detecting light scattered from droplets formed through particle-induced nucleation in a supersaturated atmosphere [33]. A schematic of the ES-DMA analysis process is presented in Fig. 1A and further discussed in Methods: ES-DMA. High resolution ES-DMA identifies multi-modal distribution of species in commercial AAV preparations. A Schematic of ES-DMA analysis. Sample loading: Analyte is loaded into the emitting end of silica capillary. Bipolar electrospray: Positive and negative emitters generate a charged aerosol in the presence of humidified air. Charge-reduced polydisperse aerosol is carried into the DMA inlet for mobility separation. Mobility selected aerosol is directed to the CPC detector. B DMA spectra of commercial “full” rAAV9 sample in units of voltage vs counts/s showing multiple repeating features, the most prominent peak of each is highlighted with a square and assigned charge state indicated above. Inset shows detail of B from 0 to -2kV. C Singly charged region of the spectra depicted in B represented in mobility diameter units showing overlay of rAAV9 “full” sample with certified empty rAAV9 reference standard. Several groups have previously applied ES-DMA to characterize the size [34], thermal decomposition [35], pH-dependent structural transitions [36], and even DNA content [37] of viral particles. However, limitations in instrument resolution have limited precise study of capsid cargo composition [38]. Foundational ES-DMA studies of AAVs from the group of Günter Allmaier reported largely overlapping spectra for two independent rAAV8 preparations enriched for empty or full capsids [39]. While a ~ 1 nm bulk difference was observed between these samples, distinct features were not sufficiently resolved for quantification. Here we employ a novel benchtop high resolution ES-DMA platform NanoRanger (NanoEngineering Corp., New Haven CT) to characterize various AAV preparations. ES-DMA analysis was performed in parallel with CD-MS to benchmark the ability of ES-DMA to quantify AAV capsid content. Native electrospray mass spectrometry has previously been used to assess stoichiometry and cargo loading of large viral and virus-like particles [40], including AAVs [41, 42]. Mobility and mass spectrometry analysis of large protein assemblies has been performed previously either in tandem (DMA-MS) to infer gas-phase structural transitions [43] or in parallel to relate electrophoretic mobility diameter (EMD) to molecular weight of virus like particles (VLPs) [44], as performed here for rAAV. We perform this side-by-side mobility and mass analysis for 12x rAAV serotype 9 (AAV9) samples nominally containing genomic cargos of various lengths from 0 - 5.0 kilobases (kb). Our results demonstrate the utility of high-resolution ES-DMA technology for AAV capsid analytics. Results To assess whether ES-DMA can quantify AAV capsid DNA content we first evaluated a commercial “full” preparation of rAAV9 nominally loaded with a ~ 5.0 kb transgene (CMV7-spCas9). The resulting spectra (Fig. 1B) shows repeating prominent peaks, highlighted with squares, at voltages corresponding to 1/n (1/2, 1/3, 1/4…) ratios, identifying them as singly, doubly, triply, etc. charged species. Supplementary Fig. SI1 plots peak voltage against charge for these peaks, showing less than 1% variation between the observed peak voltage and its expected mobility, assuming a linear charge-mobility relationship. This data unambiguously identifies the singly charged portion of the spectra (in this case from −4.25 to −5.8 kV) which represents the primary region of interest for all rAAVs analyzed in this study. The singly charged rAAV9 particles exhibit a multi-modal distribution, depicted with respect to DMA voltage (Fig. 1B, blue fill) or in electrophoretic mobility diameter (EMD) units (Fig. 1C, blue trace). (See Supporting Information Section 2 and Figure. SI2 for discussion of the electrical mobility/particle diameter relationship). This distribution is consistent with a mixture of empty/partial / full AAV capsids present within the sample. To examine this possibility, we next compared these results against an rAAV9 certified empty reference standard, which was indeed observed to produce a single peak at ~23.4 nm (Fig. 1C, red trace). Overlaying these spectra shows alignment of the smallest EMD peak of the “full” sample with that of the empty reference standard, leaving the largest EMD peak putatively corresponding to the full particle. These results led us to tentatively identify this particular mobility diameter with empty AAV particles, and subsequent increasing diameters with cargo-containing virions. These observations are consistent with a model in which, upon desolvation during electrospray ionization, AAV capsids undergo structural compaction around encapsidated cargo to produce electrical mobilities determined by DNA content. To confirm the relationship between rAAV capsid DNA content and particle size as quantified by high resolution ES-DMA we next performed parallel mass analysis using charge detection mass spectrometry (CD-MS) for a panel of rAAV9 samples nominally containing DNA cargos of varying lengths (Supporting Information Table SI1). For this panel all rAAV samples were fixed in 2% paraformaldehyde (PFA) prior to analysis. To evaluate potential artifacts introduced by fixation, additional tests were performed using samples prepared without PFA pre-treatment. These samples produced virtually identical spectra with the mass-diameter correlation for the various features being preserved despite a small peak displacement of about 0.5% in mass and diameter, presumably due to the presence of formaldehyde adducts in the fixed samples (Fig. SI3). Genome quantification was also found to be similar for both preparation methods, enabling analytics to be performed at biosafety level 1 (BSL1) facilities. Figure 2 compares CD-MS and ES-DMA spectra in parallel for a subset of these rAAV9 samples, revealing a clear correspondence between mass and mobility profiles (remaining set shown in Fig. SI4). For both CD-MS and DMA, the spectra are modeled as a set of 4-6 Gaussian curves identifying the various peaks (See Methods: Data Analysis, Fig. SI5). Assigned empty and full peaks for mass spectra match closely the expected sequence mass values for empty capsids and rAAVs loaded with the nominal genomic cargos (Fig. SI4 and Supplementary Table SI1). Taken together, parallel mass analysis demonstrates that particle size characterization by high resolution ES-DMA can produce peak profiles closely resembling mass spectra for rAAVs, albeit at a moderately lower resolution. ES-DMA comparison with CD-MS reveals size/mass correlation for AAV9 capsids containing various length DNA cargo. For each row, left and right correspond with CD-MS and DMA spectra respectively. The nominal cargo length for each is indicated at right in kilobases (kb). To quantify the correlation of mass and mobility diameter of the various prominent features identified in each spectra the corresponding pairs of mass and diameter peaks for each sample are plotted in Fig. 3A. Peak classification as “empty/partial / nominal / overfill” is assigned based on central peak mass, whether it falls below, around or over the expected mass for the nominal DNA cargo according to Table SI1, or near the mass for empty rAAV9 capsids, found to be ~3.78 MDa for these preparations. Figure 3A highlights a striking near linear mass-diameter correlation for DNA-containing rAAVs (R2 = 0.992), where an expression of the form: Mass–Diameter comparison reveals correlation between ES-DMA and CD-MS analysis. A Relationship between CD-MS mass and ES-DMA diameter for the various peaks identified in SI5 with a linear fit highlighting the correlation (R² = 0.992). This correlation is also displayed in units of density. B Comparison of CD-MS and ES-DMA relative quantification of species by peak area. C Quantification of FWHM peak resolution by ES-DMA and CD-MS where diameter resolutions FWHMD−1 have been scaled by 1/3 to approximate mass resolution for comparison. can faithfully relate mobility diameter (d) to particle mass (m). Similarly, for cargo-laden AAVs the mobility diameter (d) can be directly related to encapsidated DNA length in kilobases (c): With regards to quantification, peak areas are used to compute the relative abundance of the various species (i.e. empty, partial, nominal, overfill) for each sample in the panel. The abundance of each species, displayed in Fig. 3B as a percentage of the total, show a strong agreement between CD-MS and DMA quantification. Fittings for the various features also reveal differences in peak width between CD-MS and ES-DMA measurements (Fig. 3C). This parameter is particularly relevant as it highlights the limit at which additional species can be detected in an rAAV sample. For example, for a given rAAV preparation nominally loaded to capacity (~4.7 kb), the limit for clearly resolving genome truncations using CD-MS and ES-DMA can be estimated at roughly ~350 nt [45] and ~500 nt, respectively (for DMA see Fig. SI7). Moreover, particle resolution measured by ES-DMA also provides some structural information. For example in Fig. 3C at around the 3.7 MDa region, empty particles are observed with a noticeably larger mass resolution (RM = ~ 26) than that of scaled diameter (RD/3 = ~ 18). This behavior can likely be attributed to an incomplete compaction of the capsid shell structure for empty particles upon gas-phase transition, resulting in a multitude of aerodynamic cross sections for a relatively monodisperse mass distribution. Within our studies, the AAV9 empty reference standard exhibits two barely resolvable charge populations at approximately ~157 and ~167 charges (Fig. SI8). This feature is less pronounced but qualitatively similar to bi-disperse charge state features reported previously for empty AAV8 particles by CD-MS, which were suggested to represent distinct compaction states of the capsid [41]. Notably however, samples in the present study were chemically crosslinked in 2% paraformaldehyde prior to analysis, precluding a direct comparison of charge state distributions from previous CD-MS datasets. As a simple point of comparison between CD-MS charge state measurements and ES-DMA mobility diameters, Rayleigh limit based diameters can be considered. For example, the AAV sample with a nominal 5.0 kb cargo shown in Fig. 2 (bottom), the CD-MS charge state distribution (Fig. SI9) exhibits an upward diagonal trend relating mass and charge. Restricting the analysis to the modal empty and full particle populations at approximately ~155 and ~177 charges yields Rayleigh limit equivalent diameters of 23.05 and 25.18 nm, respectively, compared with ES-DMA mobility diameters of 23.67 and 25.78 nm. Notably however, a subset of fully loaded ( ~ 5.5 MDa) AAV particles exhibit charge states as low as ~145 charges, corresponding to an equivalent diameter of ~22 nm, well below those observed for the empty particles, highlighting the limitations of charge based diameter interpretation. Moreover, the upward mass-charge correlation is not consistently observed across all samples studied. Discussion Recombinant adeno-associated virus (rAAV) vectors represent a leading platform for gene therapy owing to their favorable safety record, broad tissue tropism, and capacity to achieve long-term transgene expression. However, challenges in rAAV manufacturing and analytical characterization continue to pose practical constraints on broader clinical translation. A persistent concern in rAAV production is the occurrence of heterogeneous capsid populations containing variable DNA content, even within highly purified clinical-grade preparations [20]. These empty, partial, and over-filled capsids reduce therapeutic potency, distort dose quantification, and contribute to increased immunogenicity [8]. Several analytical methods have been applied to address this issue, including electron microscopy (EM), analytical ultracentrifugation (AUC), mass photometry, and charge detection mass spectrometry (CD-MS), yet each technique has its own limitations in either sensitivity, resolution, throughput, availability, or cost [15]. Consequently, the application of multiple assays in parallel has become standard practice for comprehensive characterization of rAAV products [8]. In this context, we present high resolution electrospray differential mobility analysis (ES-DMA) as a promising new analytical modality, capable of resolving capsid species based on gas-phase electrical mobility —a novel and orthogonal parameter for AAV analytics corresponding to particle size. By performing side-by-side testing with CD-MS, we demonstrate that electrical mobilities quantified by ES-DMA are directly correlated with particle mass for aerosolized rAAVs across a range of genome sizes, enabling robust identification and relative quantification of empty, full, partial, and over-filled capsids. We thereby establish a quantitative relationship between rAAV DNA cargo length and particle electrophoretic mobility diameter, validating the application of ES-DMA for assessing capsid content in a benchtop format. This mass–diameter relationship is well-approximated by a linear fit, likely because the relevant mobility diameter range ( ~ 23.3 – 25.6 nm) is too narrow for significant deviation from the reasonably expected m∝d3 scaling. Notably, empty AAV particles deviate from the otherwise near-perfect linear relationship observed for DNA-containing species. This is presumably due to incomplete gas-phase compaction, additionally highlighting the ability of high-resolution ES-DMA to reveal structural insights into conformational heterogeneity. It is noteworthy that for empty AAV particles, clear bi-disperse charge state distributions have previously been observed using CD-MS, and it has been suggested that these populations may correspond to distinct compaction states of the empty capsids [41]. Such features were not clearly baseline resolved here in either the CD-MS data or ES-DMA electrical mobilities, potentially due to sample fixation and ES-DMA operating conditions. In Fig. 3B it is notable that for peak areas around 50% abundance by CD-MS, DMA quantifications tend toward slightly increased estimates. This effect can be attributed to low abundance intermediate species being included in the “full” peak by ES-DMA due in part to its lower resolving power relative to CD-MS and also due to a limitation of the applied Gaussian-based quantification model. Indeed, CD-MS has previously demonstrated capacity to resolve sub-350 nucleotide (nt) genome truncations [45], while in this study ES-DMA was able to partially resolve ~900 nt deviations from nominal cargo loading (see 2.8 kb sample Fig. SI4). However, due to the highly monodisperse peak profile of capacity-loaded rAAV capsids, any unusually broad “full” peaks identified by ES-DMA may also be used to flag rAAV batches for possible truncations, likely down to the order of ~300-500 nt, even without complete separation of peaks (see model Fig. SI7 where various cases are illustrated and resolution limits hypothesized). Such samples may then be subjected to more detailed characterization by other methods such as CD-MS or AUC. Additionally, by performing separations at ambient pressure, ES-DMA has the unique capability to collect size-separated particles at the instrument outlet for orthogonal testing via qPCR, electron microscopy or DNA sequencing. Ongoing efforts aim to exploit this capacity to characterize DNA sequence information of partial / over-filled capsids to further aid researchers in eliminating these impurities during production to reduce cost and improve safety of AAV-based therapeutics. Overall, these results highlight high resolution ES-DMA as a potentially powerful addition to the suite of tools available for researchers and production teams seeking to assess and improve the quality of rAAV preparations. Beyond AAV, we anticipate that wider access to sub-nanometer sizing of complex biologics via benchtop high resolution ES-DMA systems will find broad utility in both basic research and the biopharmaceutical industry. Materials and methods Experimental All samples were fixed in 2% paraformaldehyde (PFA) in the rAAV manufacturer-supplied buffer prior to analysis by addition of 20% aqueous PFA solution (Electron Microscopy Sciences, Cat. 15713) directly to the rAAV preparation in 1:10 dilution. Samples were incubated with fixative for 15 minutes at ~25 °C prior to buffer exchange into 200 mM aqueous ammonium acetate solution (pH adjusted to ~7.6 with ammonium hydroxide). Buffer exchange was performed using 2x P6 Micro Bio-Spin columns (Bio-Rad, 7326225) applied in series and pre-equilibrated 4 times according to the manufacturer instructions. Native AAV samples were prepared for comparison by omission of PFA fixation step, with only ammonium acetate buffer exchange performed as above. For both ES-DMA and CDMS a single dataset is produced but replicate experiments were recorded and are available upon reasonable request. ES-DMA DMA Experiments were conducted using a commercial ES-DMA instrument (see diagram Fig. 1A): “NanoRanger ES-DMA System” (NanoEngineering Corp.) modified to allow for control of the relative humidity of the analysis media. All experiments were performed at atmospheric pressure, with DMA sheath flow gas temperature of 28–31 °C and a water vapor partial pressure of 1.43 kPa (corresponding to ~40% relative humidity at 27 °C in the ES chamber). Humidifying the sample aerosol currently appears critical for optimal rAAV species separation, characterization of this phenomenon will be discussed in a future publication (Perez-Lorenzo 2026). Note that bio-nanoparticle growth due to water sorption effects is to be expected when performing mobility analysis in a humidified atmosphere [46]. The instrument consists of a bipolar electrospray source [47] combined with a high-resolution Perez type DMA [34]. The instrument records detector counts/s at the classified outlet using a condensation particle counter (Aerosol Devices Series 200 Magic Particle Counter, Berkeley, CA). DMA sheath and sample flowrates are set to 73.0 and 0.35 lpm respectively. High resolution acquisitions are performed in negative mode with 200 voltage steps within −4.0 to −5.8kV ( ~ 22-27 nm) with dwell times of 4 s. Electrical mobility values are calculated using the corresponding constant for this instrument [37] and the reading from an integrated sheath gas mass flow meter (Sensirion AG, Switzerland SFM3003). Electrical mobilities are further represented in electrophoretic mobility diameter units using the conversion detailed in Supplemental Information section 2 and Fig. SI2. The sample flow consists of room air passed through a molecular sieve drier followed by a HEPA filter. A Nafion-based humidifier (Permapure, MH series, Lakewood, NJ) saturates a fraction of the gas stream until the aforementioned ~40% relative humidity is achieved. Pressure fed capillary electrospray For the positive emitter a 50 µm I.D. silica capillary (Polymicro Technologies/Molex) of length ~25 cm, was pulled under a flame and ground until reaching a tip O.D. ~20 µm. Analyte is loaded by suctioning into the emitter tip of the positive capillary (Fig. 1A) pre-filled with the same ammonium acetate solution used for sample buffer exchange. This is performed by applying light vacuum to the capillary feed line for ~20 s to draw ~1 capillary volume (<0.5 µL) from a sample vial containing the analyte. The emitting end is then introduced back into the electrospray chamber for emission with positive backpressure of ~20mBar producing a current of 300–400 nA at ~1.5 kV. A Taylor cone is visible through a camera system and monitored for stability in conjunction with the current during sampling. For the negative emitter we also use 50 µm I.D. silica capillary but pulled to smaller tip dimensions. The neutralizing emission consists solely of ~10 mM ammonium acetate solution producing a discharge of ~ −150nA at ~ −1.6 kV without forming a visible Taylor cone. CD-MS CD-MS measurements were performed on an in-house built instrument employing an electrostatic linear ion trap with an embedded detection cylinder (Megadalton Solutions Inc. Bloomington, IN). Samples were introduced by nano-electrospray ionization (Triversa Nanomate, Advion Interchim Scientific, Ithaca, NY, USA) after buffer exchange into a volatile salt. CD-MS measurements were performed at an ion energy of 130 eV/z and a trapping time of 104.6 ms to yield a charge uncertainty of ~0.8 e (elementary charges). All datasets were processed using the MICE algorithm to enable accurate, high-throughput quantitation [48]. About 5000 ions are collected in the AAV capsid range. Data analysis The various Gaussian fittings represent the observed features and are used to obtain central mass and diameter values, as well as peak height and standard deviation in both CD-MS and DMA spectra. These results are provided in Table 2 in Supporting information. Clearly discernible narrow peaks of significant signal intensities in both CD-MS and DMA spectra are used for the correlation. Mass and mobility peaks are then simply correlated by ordinal ranking. The result of this fitting is overlaid with the original data in Fig. SI5. Supplementary information Supplemental information Table SI2 Acknowledgements We would like to Acknowledge the contributions of XD Shao, Vector Biolabs, MegaDalton Solutions, Juan Fernandez de la Mora, Bob Dana, and MakeHaven (New Haven, CT). Author contributions PD: Methodology, Investigation, Writing – Review & Editing, Original Draft Preparation. Funding Acquisition. LMY: Investigation, Review & Editing. BED: Investigation, Review & Editing. MJ: Conceptualization, Resources. JS: Funding Acquisition, Project Administration. LJP-L: Conceptualization, Supervision, Data Analysis, Writing – Review & Editing, Visualization. Funding Acquisition. Funding This work has been partially funded by NCATS award number 1R43TR005330-01A1. Data availability Raw data files used in this study are available from LJP-L upon reasonable request. Competing interests PD, JS and LJP-L are employees of NanoEngineering Corporation, which is involved in the commercialization of the ES-DMA instrument used in this study. Megadalton Solutions has an interest in providing Charged Detection Mass Spectrometry (CD-MS) commercialization. Ethical approval Ethical approval was not required for this study, as the research exclusively involves analysis of viral vectors. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 6/15/2026 Redundant NameYear citation style is removed. The article uses numbered citation style. Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41434-026-00626-0. References Associated Data Supplementary Materials Supplemental information Table SI2 Data Availability Statement Raw data files used in this study are available from LJP-L upon reasonable request.