Dataset Information

Expression profiles of lineage-depleted (Lin-) cell and mono-nucleated cell (MNC) samples derived from human umbilical cord blood

ABSTRACT: The cellular composition of heterogeneous samples can be predicted from reference gene expression profiles that represent the homogeneous, constituent populations of the heterogeneous samples. However, existing methods fail when the reference profiles are not representative of the constituent populations. We developed PERT, a new probabilistic expression deconvolution method, to address this limitation. PERT was used to deconvolve cellular composition of variably sourced and treated heterogeneous human blood samples. Our results indicate that even after correcting batch effects, cells presenting the same cell surface antigens display different transcriptional programs when they are uncultured versus culture-derived. Given gene expression profiles of culture-derived heterogeneous samples and profiles of uncultured reference populations, PERT was able to accurately recover proportions of pure populations composing the heterogeneous samples. We anticipate that PERT will be widely applicable to expression deconvolution problems using profiles from reference populations that vary from the corresponding constituent populations in cellular state but not cellular identity. Human umbilical cord blood-derived lineage negative cells and mononucleated cells

ORGANISM(S): Homo sapiens

PROVIDER: GSE40829 | GEO | 2012/09/13

SECONDARY ACCESSION(S): PRJNA175056

REPOSITORIES: GEO

ACCESS DATA

Similar Datasets

Project description:To date, most epigenetic datasets have been generated on DNA samples isolated from bulk tissues. As the proportion of individual cell types within a sample can vary across individuals, systematic differences in cellular proportions that correlate with the phenotype of interest may manifest as differences in the overall epigenetic profile. Deconvolution algorithms calculate a series of continuous variables reflecting the underlying cellular heterogeneity of each sample from the bulk tissue profile that can be used to adjust for confounding in epigenome-wide association analyses. Here we provide novel reference profiles for brain cell types for use with reference based deconvolution algorithms. We used a FANS protocol recently described by our group to purify nuclei populations from prefrontal cortex tissue from 43 adult donors. Our initial gating strategy used an antibody against NeuN to isolate neuronal nuclei in combination with an antibody against SOX10 to separate oligodendrocyte nuclei from other glial nuclei. Subsequently, in a second gating strategy we additionally included an antibody against IRF8 to enrich microglia from the NeuNNeg/SOX10Neg fraction. Our third gating strategy used an antibody against SATB2 in place of NeuN to isolate excitatory neurons. We generated DNA methylation profiles using the Illumina EPIC array for NeuNPos (neuron-enriched; n = 28), NeuNNeg/SOX10Pos (oligodendrocyte-enriched; n = 24), NeuNNeg/SOX10Neg (microglia- and astrocyte-enriched; n = 21), NeuNNeg/SOX10Neg/IRF8Pos (microglia-enriched; n = 17), NeuNNeg/SOX10Neg/IRF8Neg, (astrocyte-enriched; n = 7), SATB2Pos (excitatory neuron-enriched; n = 9), and SATB2Neg (inhibitory neuron- and glial- enriched; n = 6) nuclei populations. We have demonstrated that these are applicable for use with established deconvolution algorithms to quantify the cellular heterogeneity of the cortex and other regions of the human brain from bulk DNAm data. These variables will be critical covariates to include in future epigenetic studies of brain disorders to minimise the risk of false positive associations and improve our understanding of the changes in the brain that underpin the development of psychiatric disorders and neurodegenerative diseases.

Dataset Information

Expression profiles of lineage-depleted (Lin-) cell and mono-nucleated cell (MNC) samples derived from human umbilical cord blood

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure