Platelet size and count are heritable, quantifiable outcomes of hematopoiesis. A better understanding of genetic mechanisms that regulate hematopoiesis could yield insights into blood disorders and augment clinical translational efforts to generate platelets in vitro. For example, induced pluripotent stem cell-derived blood cells could someday ameliorate adverse effects and increase transfusable platelet supplies. However, these methods remain inefficient.

Genome wide association studies (GWAS) have linked hundreds of DNA loci with altered human platelet traits, but little is known about how these GWAS loci mechanistically impact hematopoiesis, megakaryocyte or platelet biology. Using publicly available genome-wide association and epigenetic data sets, we applied a machine-learning framework to identify epigenetic features enriched at established platelet trait association sites. From these results, we derived a quantitative prediction model that identified hematopoiesis-, megakaryocyte-, and platelet-relevant genomic loci and related genes more accurately than any prior computational method. In addition to specifying exact genetic variants known to regulate platelet traits and function, our model highlighted several novel variants in established platelet trait variation GWAS loci that were predicted to disrupt key transcription factor binding sites.

Among loci nominated by our statistical model was a variant (rs11071720) that alters Tropomyosin 1 (TPM1) gene expression. TPM1 regulates cytoskeletal biology in many cell types, and cytoskeletal functions critically impact hematopoiesis and terminal blood cell biology. We used CRISPR/Cas9-mediated genome editing to create TPM1-knockout human induced pluripotent stem cells. TPM1KO cells were healthy and showed normal developmental progression through primitive streak, mesoderm, and early primitive hematopoietic differentiation programs. Unexpectedly, TPM1KO cells demonstrated enhanced formation of hemogenic endothelium and hematopoietic progenitor cells, increasing total megakaryocyte yield by more than 2-fold. TPM1KO megakaryocytes demonstrated normal morphology, gene expression patterns and functional responses to platelet agonists.

Our findings help explain human genetics associations and identify a novel strategy to enhance in vitro hematopoiesis. More broadly, our now-validated machine learning-based approach might specify additional functional loci underlying hematopoiesis, megakaryopoiesis, or thrombopoiesis.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.