Back

Single cell long-read genotyping of transcriptomes reveals discrete mechanisms of clonal evolution in post-myeloproliferative neoplasm acute myeloid leukemia.

Grabek, J.; Straube, J.; Cooper, L.; Haldar, R.; Zhang, R.; Dulige, I.; Barker, M.; Gatehouse, W.; Christensen, H.; Amor, G.; Ling, V. Y.; McNamara, C.; Perkins, A.; Ross, D.; Bywater, M. J.; Lane, S. W.

2025-08-22 cancer biology
10.1101/2025.08.18.670417 bioRxiv
Show abstract

Myeloproliferative neoplasms (MPNs) are caused by acquired mutations in hematopoietic stem and progenitor cells (HSPCs). The acquisition of additional mutations like TP53 and the overall mutational burden influence a patients risk of disease progression toward lethal post-MPN acute myeloid leukemia (AML). Recent technological advancements in linking single-cell gene expression with genotype have improved our understanding of tumor heterogeneity. However, current methodologies have limitations in simultaneously genotyping low-expression genes (such as JAK2) alongside other pathogenic loci. To address this, we developed a novel long read genotyping pipeline of cDNA transcripts called LOTR-Seq, which can genotype the full length of expressed transcripts of 30 genes at once. Using LOTR-Seq, we genotyped HSPCs at the JAK2V617 locus in 9,075 single cells from eight patients with chronic phase MPN (CP-MPN) and in 5,016 cells from four patients with post-MPN AML. We then linked the mutations to the single cell transcriptome of 29,712 JAK2V617F-driven CP-MPN cells and 16,895 post-MPN AML cells. In our analysis of post-MPN AMLs, we identified nine mutated loci across six genes (JAK2, IDH1/2, TP53, SRSF2, U2AF1) and linked these mutations to specific transcriptional phenotypes. Overall, LOTR-Seq provides novel insights into the evolution of post-MPN AML.

Published in Blood Advances (predicted rank #3) · training set

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.