#Deconvolution
New paper from the group describing a few new artifact suppression features in UniDec:

pubs.acs.org/jamsef/artic...

Try out the new features in the latest version of UniDec: github.com/michaelmarty...
Revisiting Artifact Suppression in Electrospray Ionization Deconvolution
Abstract. Charge-state assignment is a critical step in many workflows for intact protein analysis by mass spectrometry. Computational deconvolution algori
pubs.acs.org
November 16, 2026 at 2:59 PM
(JASMS) Revisiting Artifact
Suppression in Electrospray Ionization
Deconvolution: AbstractCharge-state
assignment is a critical step in many workflows
for
intact protein analysis by mass spectrometry. Computational deconvolution
algorithms are increasingly used to automate data… (RSS) #MassSpecRSS
Revisiting Artifact Suppression in Electrospray Ionization Deconvolution
AbstractCharge-state assignment is a critical step in many workflows for intact protein analysis by mass spectrometry. Computational deconvolution algorithms are increasingly used to automate data analysis, but conventional approaches can introduce artifacts. Here, we explore several approaches to removing artifacts by limiting the charge-state assignments for each data point. One approach removes satellite charge states surrounding the local maxima. Another removes harmonics at double the more probable charge assignment. The last two limit either the total number of allowed charge-state assignments or the minimum relative charge-state assignment probability for a given data point. These new features are as effective as the prior SoftMax artifact removal in UniDec, but they enable easier tuning by the user. Although the new set of features introduces additional complexity and an expanded number of parameters, these tools enable users to more precisely reduce artifacts in the deconvolution results.
dlvr.it
November 15, 2026 at 9:05 PM
Open Access UCL Research: A Practical Approach to FMCW Radar Deconvolution in the Sea Ice Domain discovery.ucl.ac.uk/id/eprint/10...
A Practical Approach to FMCW Radar Deconvolution in the Sea Ice Domain - UCL Discovery
UCL Discovery is UCL's open access repository, showcasing and providing access to UCL research outputs from all UCL disciplines.
discovery.ucl.ac.uk
November 18, 2026 at 12:16 PM
SpOTr combines spot deconvolution with Fused Gromov–Wasserstein transport. Separate perturbations of the tree and deconvolution inputs would show which errors drive clone-to-spot assignments; a tree-removal ablation would isolate the value of lineage information.
November 14, 2026 at 7:50 PM
A deep learning emulator for single-cell informed deconvolution and state prediction of cancer cells in clinical cohorts #NeuroDegeneration 🧪🧠
https://www.researchsquare.com/article/rs-11726977/latest
November 16, 2026 at 5:02 PM
The processing pipeline in SVI Huygens: Stitching, multiview fusion, deconvolution, and MIP-rendering svi.nl/Huygens-Ligh...
Huygens Light Sheet Software | Scientific Volume Imaging
Scientific Volume Imaging to provides reliable, high quality, easy to use image processing tools for scientists working in light microscopy. Together with a dedicated team in close contact with the in...
svi.nl
November 17, 2026 at 2:39 PM
SpaCoDA: fast compositional deconvolution maps recurrent cellular niches across hundreds of spatial transcriptomics sections #SingleCell 🧪🧬🖥️
https://www.researchsquare.com/article/rs-11482224/latest
November 13, 2026 at 8:01 AM
G. B. Scharmer: Improved compensation for missing high-order aberrations with multi-frame blind deconvolution and phase diversity techniques https://arxiv.org/abs/2611.22753 https://arxiv.org/pdf/2611.22753 https://arxiv.org/html/2611.22753
November 18, 2026 at 6:51 AM
Eight modules cover gene set and pathway visualization, ligand–receptor colocalization, cell-type deconvolution, differential expression, and multi-sample ROI comparison, with ROIs exportable as Seurat objects or reusable spot-index files.
November 17, 2026 at 11:03 AM
A deep learning emulator for single-cell informed deconvolution and state prediction of cancer cells in clinical cohorts #SingleCell 🧪🧬🖥️
https://www.researchsquare.com/article/rs-11726977/latest
November 16, 2026 at 5:01 PM
The authors included RRIDs in their in STAR Protocols paper! We value the author's support of reproducibility. #STMpublishing #ReproducibleResearch #reproducibility
Protocol for reference-free single-cell-resolution deconvolution of spot-based spatial transcriptomics with CoexpressDeconvolve
Read the full paper: Protocol for reference-free single-cell-resolution deconvolution of spot-based spatial transcriptomics with CoexpressDeconvolve
doi.org
November 15, 2026 at 8:01 AM
(ChemRxiv) Deconvolution for high-resolution oversampling MALDI MSI: Decreasing pixel size is essential to resolve single cells via mass spectrometry imaging. Here, we present a bioanalytical and computational pipeline to perform oversampling MALDI-MSI from tissues and cells and… #MassSpecRSS
Deconvolution for high-resolution oversampling MALDI MSI
Decreasing pixel size is essential to resolve single cells via mass spectrometry imaging. Here, we present a bioanalytical and computational pipeline to perform oversampling MALDI-MSI from tissues and cells and deconvolve oversampled MALDI-MSI results. By comparing to histochemical staining and fluorescence microscopy, we can show that deconvolution improves the quality of resulting images compared to unprocessed MALDI-MSI data. Despite the signal loss due to oversampling compared to regular MALDI-MSI data acquisition, we can obtain 166 annotations in positive-ion mode for mouse kidney. At the same time, we show that we can visualize single cells in tissue with a pixel resolution of 600 nm, subcellular distributions in cells with down to 900 nm, and enable measurements of different tissues and cells in both ion modes. By using mouse spleen tissue and combined fluorescence microscopy, we demonstrate that we can visualize single neutrophils within spleen tissue and obtain lipid profiles for these tissue-resident single innate immune cells.
dlvr.it
November 16, 2026 at 2:13 PM
Deconvolution helps, but near saturation it’s ill-posed and can manufacture structure. I’d forward-model the instrument response, then propagate its uncertainty into the β-vectors. A detection boundary should survive posterior draws—not appear only in one corrected trace.
November 17, 2026 at 12:16 AM
Dead time also discards arrivals, so pointwise deconvolution can’t recover them uniquely. I’d fit a latent burst process through the calibrated detector response—dead time, pile-up, dark counts—then propagate posterior uncertainty into the β-vectors. Stable features deserve biological tests.
November 16, 2026 at 11:47 PM
InSAR Fringe-Order LiDAR–Photometric Depth Fusion
Acoustic Time-Reversal Mirrors × Frankot-Chellappa Poisson
Högbom CLEAN Deconvolution × NIPR Point-Cloud Purification
Dual-Phase Lock-In Amplifier × 120Hz DET-CLOCK Photometric Stereo

vibe coded <3
November 14, 2026 at 7:39 PM
Integrated multi-omics of 214 pediatric adrenocortical tumors identifies four risk groups, linking CpG hypermethylation and WNT/AP-1 signaling to high-risk disease.

https://onkogene.com/en/article/pmid_42736301/
#CancerGenomics #MolecularOncology
November 16, 2026 at 8:00 AM
For non-Gaussian deconvolution of $\lambda(t)$, investigate Bayesian Non-Parametrics (BNP) like Dirichlet Process Mixture Models (DPMMs). They estimate intrinsic distribution flexibly without assuming fixed forms, helping robustly separate detector noise from true physical kinetics. 🧠📊
November 14, 2026 at 9:43 AM
📝 DeMixNB: deconvolution of sparse-count RNA sequencing data for tumor cells
✍️ RNA-Seq Blog
DeMixNB: deconvolution of sparse-count RNA sequencing data for tumor cells
Article de RNA-Seq Blog
www.rna-seqblog.com
November 11, 2026 at 8:40 AM
Primarily biological kinetics—but afterpulsing can masquerade as a fast tail. I’d avoid single-kernel deconvolution: map the PMT with dark runs and attenuated LED pulses, then fit instrument response and burst kinetics jointly.
November 12, 2026 at 11:16 PM
Right — peak rate and duration get convolved with the detector's own response, so what you measure is a filtered burst, not the source. Deconvolution matters if you want to compare across labs. Are you thinking of PMT afterpulsing or the biological rise/fall kinetics?
November 12, 2026 at 8:40 PM
photon-mosaic v0.0.1 is out!

A modular Python package for calcium imaging. The full suite2p pipeline with @spikeinterface.bsky.social design: easy API, provenance, laziness.

Built with @alleninstitute.org & @intlbrainlab.bsky.social.

See photon-mosaic.org for more details.
November 10, 2026 at 9:50 AM
(JASMS) Two-Dimensional Correlation
Mass Spectrometry for
Precursor/Fragment Ion Identification of Composite and Unresolved
Spectral Peaks: AbstractTwo-dimensional correlation mass spectrometry (2D-CMS)
is a cross-correlation-based
technique that can be used to identify… (RSS) #MassSpecRSS
Two-Dimensional Correlation Mass Spectrometry for Precursor/Fragment Ion Identification of Composite and Unresolved Spectral Peaks
AbstractTwo-dimensional correlation mass spectrometry (2D-CMS) is a cross-correlation-based technique that can be used to identify mass-spectral signals that arise from a common molecular origin. In this study, 2D-CMS is explored for mixtures of the blood-pressure medications atenolol and metoprolol with the specific aims of precursor/fragment-ion identification and deconvolution of unresolved spectral peaks. Atenolol and metoprolol are structurally similar and produce several identical or isobaric fragment ions. In addition, the first isotopic peak of the atenolol ion (i.e., [MH++1]) at m/z 268.1737 and the protonated metoprolol ion at m/z 268.1834 produce an isobaric overlap for the mass-spectral resolving power used in this study (m/Δm = 17,500 at m/z 200). These phenomena result in numerous peaks that originate from both analytes, complicating the interpretation of mass-spectral results for precursor/fragment identification and assignment of overlapped peaks. Semiquantitative information extracted from 2D-CMS results enabled the determination of the relative contribution of each analyte to the total signal of their shared fragments. In addition, narrow m/z regions within unresolved mass-spectral peaks could be attributed to their specific molecular origin. For a series of solutions with varying proportions of each analyte, metrics derived from 2D-CMS were found to scale with the relative concentration of each analyte. Calibration results obtained from 2D-CMS were comparable to those obtained from traditional peak-fitting methods but provided additional information that enabled peak identification. These results show the utility of 2D-CMS for deconvolution of mass spectra acquired without separation or isolation of analytes or at low resolving powers.
dlvr.it
November 14, 2026 at 8:04 PM
ALMA is an interferometer. It samples the sky’s spatial Fourier components across dozens of antenna pairs, and the image is synthesised from those data, typically with CLEAN-style deconvolution.

So it's reconstructed from data, not photographed. Still incredibly impressive though.
November 7, 2026 at 7:56 PM
Haley Duba-Sullivan, Emma J. Reid, Charles A. Bouman, Gregery T. Buzzard: AdaptivePnP: Fast Auto-Tuned Plug-and-Play for Deconvolution https://arxiv.org/abs/2611.09845 https://arxiv.org/pdf/2611.09845 https://arxiv.org/html/2611.09845
November 9, 2026 at 6:45 AM
🦠 New research in Bioinformatics Advances: “DeSHAPE-α: Deconvolution of quantile Structures of Heterogeneity, Asymmetry, and Pattern to advance microbial α-diversity analysis” 

Explore the paper: https://doi.org/10.1093/bioadv/vbag329
November 4, 2026 at 1:02 PM