Uncertainty quantification of reference-based cellular deconvolution algorithms

Vellame, Dorothea Seiler; Shireby, Gemma; MacCalman, Ailsa; Dempster, Emma; Burrage, Joe; Gorrie-Stone, T.J.; Schalkwyk, Leonard C.; Mill, Jonathan; Hannon, Eilís

doi:10.1080/15592294.2022.2137659

Cited by 10 publications

(14 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When comparing the performance of different reference panels we have demonstrated how our accuracy metric for cellular deconvolution, CETYGO (55), can be applied. Our results reinforce the conclusions from the original work, that the parameters of the distribution of the CETYGO score are reference panel and technology specific.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Quantifying the proportion of different cell types in the human cortex using DNA methylation profiles

Hannon

Dempster

Chioza

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Background Due to inter-individual variation in the cellular composition of the human cortex, it is essential that covariates that capture these differences are included in epigenome-wide association studies using bulk tissue. As experimentally derived cell counts are often unavailable, computational solutions have been adopted to estimate the proportion of different cell-types using DNA methylation data. Here, we validate and profile the use of an expanded reference DNA methylation dataset incorporating two neuronal- and three glial-cell subtypes for quantifying the cellular composition of the human cortex. Results We tested eight reference panels containing different combinations of neuronal- and glial-cell types and characterized their performance in deconvoluting cell proportions from computationally reconstructed or empirically-derived human cortex DNA methylation data. Our analyses demonstrate that these novel brain deconvolution models produce accurate estimates of cellular proportions from profiles generated on postnatal human cortex samples, they are not appropriate for the use in prenatal cortex or cerebellum tissue samples. Applying our models to an extensive collection of empirical datasets, we show that glial cells are twice as abundant as neuronal cells in the human cortex and identify significant associations between increased Alzheimer's disease neuropathology and the proportion of specific cell types including a decrease in NeuNNeg/SOX10Neg nuclei and an increase of NeuNNeg/SOX10Pos nuclei. Conclusions Our novel deconvolution models produce accurate estimates for cell proportions in the human cortex. These models are available as a resource to the community enabling the control of cellular heterogeneity in epigenetic studies of brain disorders performed on bulk cortex tissue.

show abstract

Section: Discussionmentioning

confidence: 99%

“…This process was repeated for 10 different train-test splits of the reference data. This methodology was implemented using functions in the CETYGO package (55) which are adaptations of functions from the minfi package (60) that takes matrices of beta values as input for the training and testing data.…”

Section: Generation Simulated Bulk Brain Profilesmentioning

confidence: 99%

Quantifying the proportion of different cell types in the human cortex using DNA methylation profiles

Hannon

Dempster

Chioza

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Similar to other types of genomic data, DNA methylation data derived from bulk tissue is susceptible to biases arising from variations in the cellular makeup. To address this issue, we utilized the recently developed R package ‘CEll TYpe deconvolution Goodness’ (CETYGO) [55, 62]. Building upon the functionalities of the deconvolution algorithm in the minfi package, CETYGO incorporates estimations of relative proportions of neurons (NeuN+), oligodendrocytes (SOX10+), and other glial brain cell types (Double−[NeuN−/SOX10−]) based on reference data obtained from fluorescence-activated sorted nuclei from cortical brain tissue [55].…”

Section: Methodsmentioning

confidence: 99%

DNA methylation patterns in the frontal lobe white matter of multiple system atrophy, Parkinson’s disease, and progressive supranuclear palsy: A cross-comparative investigation

Murthy,

Fodder,

Miki

et al. 2024

Preprint

Self Cite

View full text Add to dashboard Cite

Multiple system atrophy (MSA) is a rare neurodegenerative disease characterized by neuronal loss and gliosis, with oligodendroglial cytoplasmic inclusions (GCI’s) containing α-synuclein being the primary pathological hallmark. Clinical presentations of MSA overlap with other parkinsonian disorders such as Parkinson’s disease (PD), dementia with Lewy bodies (DLB), and progressive supranuclear palsy (PSP), posing challenges in early diagnosis. Numerous studies have reported perturbations in DNA methylation in neurodegenerative diseases, with candidate loci being identified in various parkinsonian disorders including MSA, PD, and PSP. Although MSA and PSP present with substantial white matter pathology, alterations in white matter have also been reported in PD. However, studies comparing the DNA methylation architectures of white matter in these diseases are lacking. We therefore aimed to investigate three parkinsonian diseases, MSA, PD, and PSP, to identify shared and disease-specific DNA methylation alterations in white matter. Genome-wide DNA methylation profiling of frontal lobe white matter of individuals with MSA (n=17), PD (n=17), and PSP (n=16) and controls (n=15), using the Illumina EPIC array, revealed substantial commonalities in DNA methylation perturbations in MSA, PD, and PSP. We further used weighted gene correlation network analysis to identify disease-associated co-methylation signatures and identified dysregulation in processes relating to Wnt signalling, signal transduction, endoplasmic reticulum stress, mitochondrial processes, RNA interference, and endosomal transport. Our results highlight several shared DNA methylation perturbations and pathways indicative of converging molecular mechanisms contributing towards neurodegeneration in the white matter of all three parkinsonian diseases.

show abstract

“…We calculated the CETGYO error metric ( 85 ) (a RMSE measure) for the Houseman deconvolutions (fig. S11).…”

Section: Methodsmentioning

confidence: 99%

Brain cell-type shifts in Alzheimer’s disease, autism, and schizophrenia interrogated using methylomics and genetics

Yap,

Vo,

Heffel

et al. 2024

Sci. Adv.

Self Cite

View full text Add to dashboard Cite

Few neuropsychiatric disorders have replicable biomarkers, prompting high-resolution and large-scale molecular studies. However, we still lack consensus on a more foundational question: whether quantitative shifts in cell types—the functional unit of life—contribute to neuropsychiatric disorders. Leveraging advances in human brain single-cell methylomics, we deconvolve seven major cell types using bulk DNA methylation profiling across 1270 postmortem brains, including from individuals diagnosed with Alzheimer’s disease, schizophrenia, and autism. We observe and replicate cell-type compositional shifts for Alzheimer’s disease (endothelial cell loss), autism (increased microglia), and schizophrenia (decreased oligodendrocytes), and find age- and sex-related changes. Multiple layers of evidence indicate that endothelial cell loss contributes to Alzheimer’s disease, with comparable effect size to APOE genotype among older people. Genome-wide association identified five genetic loci related to cell-type composition, involving plausible genes for the neurovascular unit ( P2RX5 and TRPV3 ) and excitatory neurons ( DPY30 and MEMO1 ). These results implicate specific cell-type shifts in the pathophysiology of neuropsychiatric disorders.

show abstract

Uncertainty quantification of reference-based cellular deconvolution algorithms

Cited by 10 publications

References 46 publications

Quantifying the proportion of different cell types in the human cortex using DNA methylation profiles

Quantifying the proportion of different cell types in the human cortex using DNA methylation profiles

DNA methylation patterns in the frontal lobe white matter of multiple system atrophy, Parkinson’s disease, and progressive supranuclear palsy: A cross-comparative investigation

Brain cell-type shifts in Alzheimer’s disease, autism, and schizophrenia interrogated using methylomics and genetics

Contact Info

Product

Resources

About