Gruffi: an algorithm for computational removal of stressed cells from brain organoid transcriptomic datasets

Abstract

Organoids enable in vitro modeling of complex developmental processes and disease pathologies. Like most 3D cultures, organoids lack sufficient oxygen supply and therefore experience cellular stress. These negative effects are particularly prominent in complex models, such as brain organoids, and can affect lineage commitment. Here, we analyze brain organoid and fetal single‐cell RNA sequencing (scRNAseq) data from published and new datasets, totaling about 190,000 cells. We identify a unique stress signature in the data from all organoid samples, but not in fetal samples. We demonstrate that cell stress is limited to a defined subpopulation of cells that is unique to organoids and does not affect neuronal specification or maturation. We have developed a computational algorithm, Gruffi, which uses granular functional filtering to identify and remove stressed cells from any organoid scRNAseq dataset in an unbiased manner. We validated our method using six additional datasets from different organoid protocols and early brains, and show its usefulness to other organoid systems including retinal organoids. Our data show that the adverse effects of cell stress can be corrected by bioinformatic analysis for improved delineation of developmental trajectories and resemblance to in vivo data.

Platform-Agnostic CellNet (PACNet) enables cross-study meta-analysis of cell fate engineering protocols

“…Ectopic organoid culture-specific signatures have precedents in other in vitro contexts as well. For example, endoplasmic reticulum stress and glycolytic stress signatures are known to appear in brain organoids which are not reflected in primary fetal tissue signatures (Bhaduri et al, 2020; Vértesy et al, 2022). Human kidney organoid culture protocols are also known to exhibit unexpected neural signatures (Howden et al, 2019; Liu et al, 2020; Wu et al, 2018b), potentially supporting an organoid culture-specific artifactual neural signature.…”

Human cerebral organoids — a new tool for clinical neurology research

“…Whether this compromises the overall utility of the model 151 or merely affects individual cells remains under debate 98 , 234 . Meta-analysis of scRNA data 235 and multi-omics data 154 suggest that in vitro conditions create an artificial stressed state in a defined set of cells while the remaining tissue remains unaffected. To improve nutrient supply, vascularized organoids have been developed with co-culture 236 , 237 , induction 189 and transplantation 238 methods, and cultured organoid slices have also been used 102 , 239 .…”

Cerebral Organoids as an Experimental Platform for Human Neurogenomics

“…Similar findings have been reported more broadly for in vitro cultured cells [ 31 ], including cultured human primary cells [ 33 ]. Recently developed bioinformatic approaches can be leveraged to regress stress signatures [ 34 ]. As new protocols and approaches are being developed to limit the effects of cell stress in culture conditions, single-cell transcriptomic studies will serve as an invaluable resource for comparing the fidelity and robustness of these protocols to the normal developing brain.…”

“…Some of the potential consequences include genomic instability and the acquisition of somatic mutations in iPS derived models [ 104 ]. To overcome these limitations, algorithmic approaches have been developed to regress gene expression signatures related to cell stress [ 34 ]. While a variety of approaches are being examined, protocols that incorporate slicing of organoids or culture at the air-liquid interface [ 16 , 105 ] have shown a substantial reduction in the level of hypoxia within organoids.…”