Constructing a Computational Workflow for the Identification of Novel Cellular and Molecular Drivers of Human Granulopoiesis

Neutrophils are crucial immune cells with complex and heterogeneous transcriptional programs. To understand the dynamic changes governing human granulopoiesis, we previously developed an ex vivo myeloid cell differentiation assay using human mobilised peripheral blood CD34⁺ cells from healthy donors and conducted single cell multiomic analysis. We constructed a multimodal atlas of human granulopoiesis that captured the full spectrum of human myeloid cells spanning from early myeloid progenitors through to neutrophil populations, thus enabling the study the genome regulatory events underlying neutrophil maturation with unprecedented resolution¹.

Our aim was to design a computational analytical workflow enabling the exploration of the cellular and molecular drivers of normal human granulopoiesis alongside the precise characterisation of the dynamic changes in the genome regulatory landscape dictating myeloid cell commitment and terminal neutrophil differentiation.

Our dataset included transcriptomic (scRNA; n=3 donors and 29,300 cells) and simultaneous profiled gene expression and chromatin accessibility (scGEX & scATAC; n=3 donors and 17,219 cells) analyses from mature neutrophils and myeloid precursors. We successfully co-embedded our data with a published reference dataset encompassing neutrophils and myeloid progenitors from heathy donors, confirming the overlap of the transcriptomic signatures. The transition from transcriptome-based to epigenome-based cluster labels improved annotation resolution, enabling precise identification of previously unexplored subtypes within neutrophil and precursor populations. Further investigation of both modalities revealed distinct chromatin conformations, but almost identical transcriptomic signatures, for clusters of band and mature neutrophils approaching the end of their lifespan. The exhausted pro-apoptotic neutrophils displayed dynamically increased chromatin compaction as evidenced by a 42.1% decline in accessible peaks detected, indicating an earlier stage towards programmed cell death commitment compared to robust neutrophils.

Next, we used our resource atlas to interrogate the underlying genome regulatory networks and characterise cell-type-specific cis regulatory elements (CREs) involved in human granulopoiesis. We coupled transcription factor (TF) and target gene identification using Scanpy with SCENIC analysis, enabling a comparative study focusing on open chromatin regions. We then linked the pre-identified cell-type-specific activities for 178 differentially present TFs to chromatin accessibility changes using ArchR and performed motif enrichment and peak-to-gene linkage identification using the integrated scGEX dataset. As a result, we were able to underpin the connection between neutrophil terminal differentiation potential and the dynamic changes in chromatin architecture.

We analysed the role of our candidate TFs by imputing regulatory single nucleotide polymorphisms (SNPs) from published genome-wide association studies (GWAS) focusing on neutrophil biology and absolute counts. This showed an increased transcriptional activity for our candidate TFs, with an overall enrichment frequency of detected regulatory SNPs, within the metamyelocyte compartment, thus further highlighting the role of this differentiation stage as a regulatory switch towards terminal neutrophil maturation.

We used reference chromatin immunoprecipitation sequencing (ChIP-seq) data from mature neutrophils and progenitors focusing on histone modification marks and CTCF sites to further annotate de novo genome-wide cell-type-specific CREs using REgulamentary². These findings were visualised using Multi-Dimensional Viewer, a comprehensive analytical interpretation tool, generating a publicly available resource atlas benefiting from the seamless interaction with this multiomic dataset.

We present a computational analytical workflow that enables large-scale multiomic data mining to study the molecular underpinnings of normal neutrophil development, by precisely characterising the dynamical changes of the regulatory landscape during human granulopoiesis and generating testable hypothesis through regulatory SNPs assessment.

1. https://doi.org/10.1182/blood-2023-182041

2. https://doi.org/10.1101/2024.05.24.595662