Thread by @rosslevinemd

OK tweeps, today is the day, our work on single cell analysis of clonal evolution and myeloid transformation is out, led by the best in the biz, @lindemilesphd and @bowman_rl
nature.com/articles/s4158…

1. First things first; thanks to everyone who helped us with this awesome story; @thedoctorisvin, Chris F. and Minal P. in @CHM, @Levinelab (Tiff!), @duruthy & @MissionBio, @SaraMeyerLab, @RaajitRampal , @kellybolton, @ahmetz, Michael Bowman, Adam Abate and team->thank you!

2. Please also see awesome paper by @DrKTakahashi and co at @NatureComms and preprint by Adam Abate on DNA/protein sc sequencing/tech dev. All together the collective work is so cool!

3. This work started with a simple question; can we decipher somatic clonal evolution from normal stem cells to CH and myeloid malignancies? What do we learn from charting this journey?

4. We did single cell DNA sequencing on 146 samples from 123 patients with CH, MPN and AML, focusing on 31 genes recurrently mutated in these diseases

5. This allowed us to delineate mutations at single clone resolution and to build maps of clonal complexity and architecture in CH, MPN and AML, which taught us a few things..

6. First, CH (with >1 mutation) is a disease of oligoclonality, with different mutations in different clones

7. By contrast, MPN and AML were diseases of complex clonal evolution, with progressive mutational acquisition, clonal branching, and emergence of dominant clones and significant subclones.

8. The majority of AML patients have 1-2 dominant clones, and these clones frequently harbor co-occurring mutations in epigenetic regulators

9. FLT3/RAS mutations were rarely in dominant clones, and were commonly subclonal

10. We then used a hidden Markov model to predict the optimal clonal trajectory in each sample, allowing us to 1) predict the likely initiating mutation and 2) show which mutational combinations cooperate to drive clonal expansion

11. Specific mutational combinations, such as co-occuring DNMT3A/IDH and NPM1c/FLT3 mutations, promote clonal expansion. These combinations are the “secret sauce” of mutational cooperativity we need to understand.

12. Taken together, these data suggest that patients with myeloid malignancies manifest as a complex ecosystem of clones which evolves over time, and that single cell sequencing gives a glimpse into this milieu not seen with conventional bulk sequencing.

13. We then used combined single cell DNA sequencing + cell surface profiling to characterize cell surface protein expression based on clone genotype. Specific subclones, esp those w/ signaling mutations were associated w/ specific immunophenotypes (Ras + CD11c expression)

14.Importantly, in 8 of 14 AML samples, the dominant clone had a distinct immunophenotype from the minor subclones, suggesting that cell surface targeting in AML may need to be clone-directed, and not just bulk disease directed

15. We also profiled serial samples showing MPN->AML transformation and pre/post FLT3 inhibitor response/resistance, giving insights into how this approach can be used to delineate clonal evolution over time.

16. This work was all made possible by the two best colleagues I could ask for, @lindemilesphd and @bowman_rl. Their work, drive, and insight, including to get critical sequencing data done and analyzed as the lab was shutting down in March, made this happen. You both rock.

17. Lots more to do in this space, and lots of groups will be doing it! Our data and code are available, and we’ll chat with anyone about what we’ve done, how we did it, and what we learne

18. The link to the sequencing data: ncbi.nlm.nih.gov/projects/gap/c…

19. Link to the code used for analysis: github.com/bowmanr/scDNA_…

20. And like a Marvel movie, if you're still reading->We promise to take this story to the next level with mechanistic/functional studies, using this work as our "blueprint" to the Death Star. AML, we're coming for you, stay tuned!!!

21. One important note; based on my experience working with Mission Bio I recently joined their SAB (noted in paper COI). Enjoyed working with them on this project and moving forwards, pls note this work is meant to illustrate power of SC-mutationl studies no matter the platform