2026 ASH®
Annual Meeting & Exposition December 12-15, 2026 ▪ New Orleans, Louisiana

Paper Details

Abstracts, authors, affiliations, and disclosures.

Saturday, December 12, 2026
05:30 p.m. - 07:30 p.m.
MCCNO - Hall F
Access, Representation, Population Health, Technology and Procedures, Imaging, Molecular Testing, Omics Technologies, Pathology, Bioinformatics, Blood Banking, Computational Biology, Emerging Technologies, Profiling, Treatment Considerations, Drug Development, Supportive Care, Therapy Sequence, Biological Therapies, Chimeric Antigen Receptor, (CAR)-T Cell Therapies, Women’s Health, Workforce, Adverse Events, Artificial Intelligence (AI), Computer Vision, Deep Learning, Generative AI, Machine Learning, Natural Language Processing, Biological Processes, Apoptosis, Genomics, Hematopoiesis, Molecular Biology, Clinical Practice (Health Services and Quality), Clinical Guidelines, Quality Improvement, Quality of Care, Systems-Based Hematology, Diseases, Bleeding and Clotting, Thromboembolism, Bone Marrow Failure Syndromes, Inherited Marrow Failure Syndromes, Paroxysmal Nocturnal Hemoglobinuria, Lymphoid Malignancies, Lymphoid Leukemias, Lymphomas, Plasma Cell Disorders, Myeloid Malignancies, Acute Myeloid Malignancies, Chronic Myeloid Malignancies, Metabolic Disorders, Iron Deficiency, Measurable Residual Disease, Non-Hodgkin Lymphoma, B Cell Lymphoma, Aggressive Lymphoma, Research, Fundamental Science, AML, APL, Translational Research, MDS, MPN, Clinical Research, Clinical Trials, Real-World Evidence, Study Population, Human, Adult, Pediatric
PAPER

2569 Mapping the RNA isoform and genotype landscape at single-cell resolution in healthy and MDS bone marrow

Elana Thieme, BS, Ruibo Zhang, BS2, Erica Arriaga-Gomez3, Stanley Lee, PhD4, Manu Setty, PhD5
1 Fred Hutch Cancer Center, Seattle, Washington, United States
2 Fred Hutch Cancer Center, Seattle, Washington, United States
3 Fred Hutch Cancer Center, Seattle, Washington, United States
4 Fred Hutch Cancer Center, Seattle, Washington, United States
5 Fred Hutch Cancer Center, Seattle, Washington, United States

Hematopoietic cell types acquire their unique form and function in part through expression of RNA isoforms. Precise regulation of the ratio between RNA isoforms is critical to hematopoiesis. For example, the master transcription factor RUNX1 produces long and short isoforms with opposing activities, and their ratio influences stem cell expansion (PMID:22130803). Despite the key role that RNA isoforms play, the dynamics of isoform composition have not been systematically characterized in healthy hematopoiesis, let alone in disease. Somatic mutations in RNA splicing factors such as SRSF2, SF3B1, and U2AF1 are common in myelodysplastic syndromes (MDS), a difficult malignancy to study because driver lesions pass from mutant stem cells to mature lineages—requiring both isoform and single-cell resolution. 

Here, we used long read single-cell RNAseq to generate an isoform resolved atlas of human bone marrow from 4 healthy and 11 splicing mutant MDS donors. The atlas has 220K cells from whole marrow and CD34+ fractions that were processed using 10x GEM-X then PacBio Kinnex. Libraries had a median read length of 907±181bp, compared to ~90-150bp for short read scRNAseq. Data were preprocessed using a custom pipeline with the tools IsoSeq, Bambu, and IsoQuant. The gene-level data co-embedded well with public short read scRNAseq, confirming that long reads reproduce established expression profiles while adding isoform resolution. Moreover, the isoform-level UMAP closely resembled the gene UMAP, even resolving CD4 activation states better than gene expression alone due to PTPRC (CD45) alternative splicing. 

Long read single-cell data are still analyzed much like bulk data, comparing isoforms between pseudo-bulked cell types rather than across a continuum of cell states. We therefore began developing a statistical framework to study continuous changes to isoform composition in the healthy dataset. Each gene’s isoform counts are modeled to vary smoothly along the cell state manifold, borrowing information from neighboring cells with a Gaussian process; a variance-component score test then selects genes whose composition shifts. This approach generated a preliminary set of genes with isoform changes, enriched for PH domain–containing genes, in which alternative splicing is known to alter subcellular localization (PMID:30404949). Our method identified multiple well-characterized genes with alternative isoforms such as RUNX1, PTPRC, FOXP1, and MEIS1. Certain long genes (e.g. TAL1) had few full-length molecules due to long read limitations and therefore were not identifiable. 

Single-cell sequencing captures whole lineages in one snapshot, resolving cell state transitions that bulk approaches miss. For example, upregulation of the short RUNX1 isoform has been shown to promote HSPC expansion, but RUNX1 isoform composition remains incompletely characterized throughout differentiation. We found that during lymphopoiesis, cells in proB-transition-preB stages expressed RUNX1 as long-short-long. This suggests that RUNX1 function may be modulated through isoform balance in pro-preB transition. 

We also intend to study isoforms in MDS. As a first step, cells in the atlas had to be genotyped with targeted sequencing. MDS donors all carried SRSF2 P95 mutations, with 23 co-mutations. Existing amplicon-based genotyping approaches would not work because some variants were too close to 5’/3’ ends. Instead, we used a custom hybrid capture approach followed by Iso-Seq that has no constraint on variant location or number of targets per reaction. Hybrid capture increased depth 7X and recovered 2.6X more mutant cells, covering 38.5% of cells per gene on average (1.9% SETBP1 to 90.5% SRSF2). WT calls were inferred from a Bayesian model whose prior is estimated across the cell state manifold, with the posterior cutoff calibrated on known WT cells from donors with a SNP in SRSF2. Ultimately, 80% of cells were genotyped for SRSF2. We recapitulated mis-splicing of known SRSF2 targets, e.g. EZH2, INTS3, CASP8. However, not all targets were robustly mis-spliced in HSCs relative to other CD34+ lineages, underscoring the need for single-cell resolution. 

This work provides a genotype-aware isoform atlas of healthy and splicing mutant hematopoiesis, built on two technological advancements: cell-state-continuous isoform analysis and hybrid capture based genotyping. These tools open a path to novel HSC and MDS biology, from disease mechanisms to therapeutic targets. 

Elana Thieme, BS: No relevant conflicts of interest to declare.