FlowCytoNews
Flow cytometry, gated weekly.
ISSUE #2 · MONDAY EDITION · THE LITERATURE · SEP 28, 2026
Welcome to the Monday edition of FlowCytoNews. Monday is The Literature: what cytometry papers and preprints went online in the last two weeks, who did the work, and the number that matters from each abstract. Wednesday brings On the Calendar (meetings, courses, deadlines) and Friday brings Industry Moves (instruments, reagents, biopharma, and who is hiring). This fortnight has an AI that turns clusters into sorter gates, a new clinical panel series, and a lot of clinical flow. Tell us what we missed — hit reply.
■ The Literature
No new OMIP appeared in Cytometry Part A in this window. The nearest thing is more interesting: a clinical counterpart to the OMIP series, below.
An AI now writes the gating strategy you can type into the sorter
Clustering finds the population; it does not give you a gate you can sort on. ConvexGating, from Vincent Friedrich (University of Leipzig) with Maren Büttner (DZNE Bonn / Helmholtz Munich) as corresponding author and Fabian Theis and Joachim Schultze among 20 co-authors, learns "interpretable gating strategies for sorting in an unbiased, data-driven manner" from clusters, generating "low-contamination strategies for both known and previously unknown cell populations, including plasmacytoid dendritic cells identified solely as CD57−CD13−CD45RA+CD123+ cells." The team derived sorting strategies for CD8+ subtypes and adipose progenitors and validated them by single-cell sequencing of the sorted cells; the method also transfers to CyTOF and CITE-seq data. No industry-affiliated authors.
▪ Nature Communications · Sep 18 · 10.1038/s41467-026-77360-z
Why it matters: Every sorter operator has been handed a UMAP and asked "can you sort that?" This is the first widely usable tool that answers with a 2D gate hierarchy and shows the sorted product is what the cluster said it was.
Cytometry Part B launches a clinical OMIP series
An editorial from Ahmad Al-Attar (Department of Pathology, University of Pittsburgh) and Kamila Czechowska (Metafora Biosystems, Paris — a cytometry-software vendor) announces "Clinical Optimized Multiparameter Immunophenotyping Panels," a new publication series in Cytometry Part B. No abstract is online, so we are running the title only; the editorial itself is the source for scope and submission rules, and it is short enough to read before you decide whether your validated diagnostic panel belongs there.
▪ Cytometry Part B · Sep 16 · 10.1002/cyto.b.70072
Why it matters: OMIPs standardized research panels; clinical labs have had no peer-reviewed equivalent for validated diagnostic panels. This is that venue, and it means your validation package can now be a citable paper.
A core facility built a free QC dashboard that reads BD, Beckman and Cytek logs together
Britt Kolen and Tamar Tak's Flow Cytometry Core Facility at Leiden University Medical Center, with LUMC's biomedical data sciences group, published FlowCytoMonitoR, "an interactive R-shiny program that does not require coding experience and allows users to analyze QC data in full detail for multiple instruments." It is "currently compatible with instruments from three different vendors (BD Biosciences, Beckman Coulter, and Cytek Biosciences) and allows visualization of all QC metrics," addressing vendor modules where "usually only a single parameter can be visualized and visualization over time is limited to 1 year or to the last baseline." The authors say it has helped their SRL in "identifying and potentially predicting problems." No industry co-authors.
▪ Cytometry Part A · Sep 26 · 10.1002/cyto.a.70065
Why it matters: If your facility runs more than one vendor's cytometers, this is the first published tool that puts every instrument's QC history on one screen without a coding requirement.
Four hematopathology groups agree on how to call an NK clone by flow
Aaron Wilk and Jean Oak (Stanford Pathology), with Sindhu Cherian, Jonathan Fromm and Afshin Shameli (University of Washington), Bartosz Grzywacz (Beth Israel Deaconess) and Min Shi (Mayo Clinic), published a best-practices framework for NK-cell analysis that integrates "(i) robust lineage definition and gating, (ii) subset-aware interpretation of CD16/CD56 patterns, (iii) inference of clonality using killer cell immunoglobulin-like receptor (KIR; CD158) repertoire restriction." They flag "the limited utility of KIR-based clonality inference in CD56bright CD16− NK cells and the potential for adaptive NK cell expansions to appear KIR-restricted," and propose "an operational reflex algorithm triggered by NK cell expansion and/or defined immunophenotypic abnormalities in screening tubes," with worked cases including NK-LGL leukemia and extranodal NK/T-cell lymphoma. No industry co-authors.
▪ Cytometry Part B · Sep 25 · 10.1002/cyto.b.70076
Why it matters: NK clonality is the call most likely to be argued at sign-out. A multi-institution reflex algorithm with named pitfalls is what you cite in your SOP.
Same blood, two methods, half the circulating tumor cells: myeloma CTC counts depend on fix/perm
Elona Saraci and Mattia D'Agostino (University of Torino) and 29 co-authors compared single-platform flow against EuroFlow Next Generation Flow (NGF) for circulating tumor plasma cells in 34 newly diagnosed, transplant-ineligible myeloma patients from the REAL MM trial (NCT03829371). "Single-platform FC detected CTCs in 29/34 patients (85.3%), while NGF detected CTCs in 31/34 patients (91.2%), including 2 cases that were negative by single-platform FC." But NGF's CTC percentage was significantly lower: "median 0.0049%, range 0%–0.377%" versus "median 0.009%, range 0%–0.59%" for single-platform (p < 0.0001). Adding fixation and permeabilization before single-platform FC "led to results that were similar to NGF." No industry co-authors.
▪ Cytometry Part B · Sep 14 · 10.1002/cyto.b.70068
Why it matters: Published CTC prognostic cut-offs (0.01%, 0.02%, 2%…) were generated with different processing. A near two-fold method effect explains why they do not transfer between labs — and tells you which step is responsible.
A DNA-methylation assay matched CLIA flow and CBC differentials within ±2%
Pauline Patzke and Brock Christensen (Geisel School of Medicine at Dartmouth) tested "methylation cytometry" — immune cell quantification from cell-type-specific DNA methylation — against CLIA- and CAP-certified clinical flow cytometry and CBC differentials in whole blood from 101 patients. Cell proportions "showed strong concordance with clinical assays (R² = 0.99; RMSE = 2.42)," with "95.4% of values falling within the limits of agreement" on Bland–Altman analysis, and "two one-sided tests of equivalence (TOST) demonstrated that the mean differences between methylation cytometry and flow cytometry were statistically equivalent within ± 2% for all cell types." In 32 patients with HIV, CD4 T-cell percentages showed "a near-zero mean difference, with 97% of observations falling within the limits of agreement." Disclosure in the paper: Christensen and co-author Lucas Salas are co-founders of Cellintec, and Christensen advises Guardant Health.
▪ Clinical Epigenetics · Sep 26 · 10.1186/s13148-026-02246-2
Why it matters: This is the paper someone will wave at your immune-monitoring service. Know its scope — proportions of major lineages in whole blood, not activation, exhaustion or MRD — before the conversation.
Also out
Velociraptor matches cells across spectral flow, CyTOF, IMC and cycIHC. Claire Cross and Jonathan Irish (Vanderbilt / CU Anschutz) extend Marker Enrichment Modeling to register cells between platforms: "median F1-measure of 0.81" between IMC and CyTOF, and a "concordance correlation coefficient of 0.99" between spectral flow and CyTOF; code at github.com/cytolab. Early view dated Sep 10, just ahead of our window. — Cytometry Part A · Sep 10 · 10.1002/cyto.a.70066
Bulky antibody conjugates undercount double-positive EVs; nanobodies fix it. Yunyun Hu and Xiaomei Yan (Xiamen University) use nano-flow cytometry to show steric hindrance between probes suppresses co-detection, then recover it with compact dyes and nanobodies; dual-nanobody PD-L1/CD47 on plasma EVs gave "an overall accuracy of 90.0% for breast cancer diagnosis." — Small Methods · Sep 22 · 10.1002/smtd.71053
Flow MRD caught 60.9% of PCR-positive childhood B-ALL samples. Hasan Fatih Çakmaklı and colleagues (Ankara University) compared their flow lab with a EuroMRD reference RQ-PCR lab in Heidelberg on 85 paired marrows from 46 children: concordance 87.1% (κ = 0.64), specificity 96.8%, sensitivity 60.9%; flow alone "would have altered MRD-guided treatment allocation in 17% of patients at EOI and 7% at EOC." — Turkish Journal of Haematology · Sep 14 · 10.4274/tjh.galenos.2026.98475
Pre-transplant flow MRD predicts outcome in pediatric AML. Barbara Buldini (University of Padova) and Franco Locatelli (Bambino Gesù, Rome) for the Italian AIEOP centres. Published as a letter with no abstract online, so title only. — Haematologica · Sep 17 · 10.3324/haematol.2026.300844
Allele-specific antibodies on the cytometer flagged HLA-loss relapse after haplo-HCT. Takumi Kimura, Yoshitaka Zaimoku and Toshihiro Miyamoto (Kanazawa University Hospital), in adult T-cell leukemia/lymphoma. A case report; no abstract online, so title only. — Haematologica · Sep 17 · 10.3324/haematol.2026.301574
Label-free basophil activation by impedance cytometry. Wonjun Lee and Sindy Tang (Stanford Mechanical Engineering) unfold a signal-processing algorithm into a deep network to demultiplex code-multiplexed, multi-frequency impedance data; "impedance opacity correlates well with activation levels measured by fluorescence flow cytometry." — Microsystems & Nanoengineering · Sep 21 · 10.1038/s41378-026-01420-z
Mixed normal and aberrant mast cells mean milder disease. Abdulrazzaq Alheraky and André Mulder (UMC Groningen), 207 patients with clonal mast cell disease: partial aberrancy in "75% of patients with MMAS, 67% with non-advanced SM, and 27% with advanced SM," one of the most discriminative parameters between indolent and advanced subtypes. — Cytometry Part B · Sep 10 · 10.1002/cyto.b.70071
A proposal for International Flow Cytometry Day on Nov 12. Editorial from Jordi Petriz and Marc Sorigue (IGTP Barcelona), Alberto Álvarez-Barrientos (Universidad de Extremadura) and José Enrique O'Connor (University of Valencia). No abstract. — Cytometry Part A · Sep 21 · 10.1002/cyto.a.70067
Preprints
Digital-PCR-grade nucleic-acid cytometry on the MACS and FACS you already own. magniFIND-seq, from Seung Won Shin and Iain Clark (UC Berkeley), captures single-cell genomes in agarose beads, detects targets by digital PCR, then labels bead-bound amplicons with magnetic or fluorescent probes; "controlled evaporation shrinks beads from 55 to 20 μm, enabling scalable magnetic separation, fast single-bead FACS sorting." They isolate SIV proviral pol and env targets and enrich BCR::ABL1+ CML cells for single-cell transcriptomes. — bioRxiv · Sep 18 · 10.64898/2026.09.12.751168
Optimal transport for sample-level classification and MRD estimation. FlowLOT, from Naqib Sad Pathan and Gustavo Rohde (University of Virginia) with Bartek Rajwa (Purdue), reaches "94.3% and 98.0% balanced accuracy" for AML vs healthy with 16 and 8 patients per class on FlowCAP-II and BLAST110, and at the 0.1% LAIP threshold "detects positivity with 72% sensitivity at 100% specificity." — arXiv 2609.17906 · Sep 15 · arxiv.org/abs/2609.17906
A vendor's own numbers for closed, automated point-of-care CAR-T. Corresponding author Caroline Boudousquie is at Limula (the instrument maker; academic co-authors are listed but their affiliations were not retrievable). Automated buoyancy-based CD3+ selection "achieved 90.3 ± 3.5% purity and 59.2 ± 17.0% recovery in under two hours"; a 7-day closed CD19 CAR-T process needed "155 minutes of operator intervention" and gave "43.6 ± 6.0% transduction efficiency" — all flow readouts. — bioRxiv · Sep 14 · 10.64898/2026.09.08.748871
■ Bench Notes
Before you compare two methods for a rare population, make the sample prep identical
This fortnight's myeloma CTC paper (Saraci et al., Cytometry B, 10.1002/cyto.b.70068) is a clean lesson: the same 34 blood samples gave a median 0.009% CTCs by single-platform flow and 0.0049% by NGF, and the difference disappeared once the single-platform samples were fixed and permeabilized. If you are validating a new rare-event method against an old one, do this:
Split every sample and run both arms from one draw, same day. Time-to-stain differences in a rare-event assay masquerade as method differences. The Torino group ran both techniques "simultaneously" on the same peripheral blood.
Match the pre-analytical chain before you compare numbers. Bulk lysis, fix/perm and wash steps each change cell loss and the denominator. Run the old method with the new method's processing (as Saraci et al. did) so you can attribute the shift to processing rather than to the panel.
Report the denominator and the acquired-event count with every percentage. A 0.0049% result on 5 × 10⁶ events is 245 cells; on 5 × 10⁵ it is 24. The EuroFlow NGF MRD guidance (Flores-Montero et al., Leukemia 2017, 10.1038/leu.2017.29) sets the ≥ 10⁷-event target that makes the 10⁻⁵ to 10⁻⁶ range meaningful; the same logic applies to CTCs.
Compare detection rate and quantity separately. Here NGF detected more positives (31/34 vs 29/34) yet reported lower percentages — one method is more sensitive, the other reads higher. Bland–Altman on the log scale shows the bias; a simple correlation hides it. The Ankara B-ALL paper above (10.4274/tjh.galenos.2026.98475) is the model: concordance, κ, sensitivity and specificity, each at each time point.
Do not import a cut-off you did not generate with the same processing. If your lab adopts a published CTC threshold, reproduce the originating lab's prep or re-derive the threshold in your own cohort. CLSI H62 (Validation of Assays Performed by Flow Cytometry, 2021) covers method-comparison design and acceptance criteria for exactly this situation.
What did we miss? Hit reply — corrections run at the top of the next edition. If a colleague runs a sorter, a clinical flow bench or an MRD assay, forward this to them.
— Josh