In sjczheng/EpiDISH: Epigenetic Dissection of Intra-Sample-Heterogeneity

Introduction

The EpiDISH package provides tools to infer the fractions of a priori known cell subtypes present in a DNA methylation (DNAm) sample representing a mixture of such cell-types. Inference proceeds via one of 3 methods (Robust Partial Correlations-RPC[@EpiDISH], Cibersort-CBS[@CBS], Constrained Projection-CP[@CP]), as determined by the user. Besides, we also provide a function - CellDMC which allows the identification of differentially methylated cell-types in Epigenome-Wide Association Studies(EWAS)[@CellDMC]. For now, the package contains 6 DNAm reference matrices, three of which are designed for whole blood [@EpiDISH] and [@MetaEWAS]:

1. centDHSbloodDMC.m: This DNAm reference matrix for blood will estimate fractions for 7 immune cell types (B-cells, NK-cells, CD4T and CD8T-cells, Monocytes, Neutrophils and Eosinophils).
2. cent12CT.m: This DNAm reference matrix for blood and EPIC-arrays will estimate fractions for 12 immune-cell types (naive and mature B-cells, naive and mature CD4T-cells, naive and mature B-cells, T-regulatory cells, NK-cells, Neutrophils, Monocytes, Eosinophils, Basophils).
3. cent12CT450k.m: This DNAm reference matrix for blood and Illumina 450k-arrays will estimate fractions for 12 immune-cell types (naive and mature B-cells, naive and mature CD4T-cells, naive and mature B-cells, T-regulatory cells, NK-cells, Neutrophils, Monocytes, Eosinophils, Basophils).

The other 3 DNAm reference matrices are designed for solid tissue-types [@HEpiDISH]:

1. centEpiFibIC.m: This DNAm reference matrix is designed for a generic solid tissue that is dominated by an epithelial, stromal and immune-cell component. It will estimate fractions for 3 broad cell-types: a generic epithelial, fibroblast and immune-cell type.
2. centBloodSub.m: This DNAm reference matrix is designed for a solid tissue-type and will estimate immune cell infiltration for 7 immune cell subtypes. This DNAm reference matrix is meant to be applied after centEpiFibIC.m to yield proportions for 7 immune cell subtypes alongside the total epithelial and total fibroblast fractions.
3. centEpiFibFatIC.m: This DNAm reference matrix is a more specialised version for breast tissue and will estimate total epithelial, fibroblast, immune-cell and fat fractions.

How to estimate cell-type fractions in blood

We show an example of using our package to estimate 7 immune cell-type fractions in whole blood. We use a subset beta value matrix of GSE42861 (detailed description in manual page of LiuDataSub.m). First, we read in the required objects:

library(EpiDISH)
data(centDHSbloodDMC.m)
data(LiuDataSub.m)

BloodFrac.m <- epidish(beta.m = LiuDataSub.m, ref.m = centDHSbloodDMC.m, method = "RPC")$estF

We can easily check the inferred fractions with boxplots. From the boxplots, we observe that just as we expected, the major cell-type in whole blood is neutrophil.

boxplot(BloodFrac.m)

If we wanted to infer fractions at a higher resolution of 12 immune cell subtypes, we would replace centDHSbloodDMC.m in the above with cent12CT450k.m because this is a 450k DNAm dataset. For an EPIC whole blood dataset, you would use cent12CT.m.

How to estimate generic cell-type fractions in a solid tissue

To illustrate how this works, we first read in a dummy beta value matrix DummyBeta.m, which contains 2000 CpGs and 10 samples, representing a solid tissue:

data(centEpiFibIC.m)
data(DummyBeta.m)

Notice that centEpiFibIC.m has 3 columns, with names of the columns as EPi, Fib and IC. We go ahead and use epidish function with RPC mode to infer the cell-type fractions.

out.l <- epidish(beta.m = DummyBeta.m, ref.m = centEpiFibIC.m, method = "RPC") 

Then, we check the output list. estF is the matrix of estimated cell-type fractions. ref is the reference centroid matrix used, and dataREF is the subset of the input data matrix over the probes defined in the reference matrix.

out.l$estF
dim(out.l$ref)
dim(out.l$dataREF)

Note: As part of the quality control step in DNAm data preprocessing, we might have to remove bad probes; consequently, not all probes in the reference matrix may be available in a given dataset. By checking ref and dataREF, we can extract the probes actually used for estimating cell-type fractions. As shown by us [@HEpiDISH], if the proportion of missing reference matrix probes is more than a third, then estimated fractions may be unreliable.

How to estimate immune cell-type fractions in a solid tissue using HEpiDISH

HEpiDISH is an iterative hierarchical procedure of EpiDISH designed for solid tissues with significant immune-cell infiltration. HEpiDISH uses two distinct DNAm references, a primary reference for the estimation of total epithelial, fibroblast and immune-cell fractions, and a separate secondary non-overlapping DNAm reference for the estimation of underlying immune cell subtype fractions. In this example, the third cell-type in the primary DNAm reference matrix is the total immune cell fraction. We would like to know the fractions of 7 immune cell subtypes, in adddition to the epithelial and fibroblast fractions. So we use a secondary reference, which contains 7 immnue cell subtypes, and let hepidish function know that the third column of primary reference should correspond to the secondary DNAm reference matrix. (We only include 3 cell-types of the centBloodSub.m reference because we mixed those three cell-types to generate the dummy beta value matrix.)

data(centBloodSub.m)
frac.m <- hepidish(beta.m = DummyBeta.m, ref1.m = centEpiFibIC.m, ref2.m = centBloodSub.m[,c(1, 2, 5)], h.CT.idx = 3, method = 'RPC')
frac.m

More info about different methods for cell-type fractions estimation

We compared CP and RPC in [@EpiDISH]. And we also published a review article[@review] which discusses most of algorithms for tackling cell heterogeneity in Epigenome-Wide Association Studies(EWAS). Refer to references section for more details.

How to identify differentially methylated cell-types in EWAS

After estimating cell-type fractions, we can then identify differentially methylated cell-types and their directions of change using CellDMC [@CellDMC]function. The workflow of CellDMC is shown below.

We use a binary phenotype vector here, with half of them representing controls and other half representing cases.

pheno.v <- rep(c(0, 1), each = 5)
celldmc.o <- CellDMC(DummyBeta.m, pheno.v, frac.m)

The DMCTs prediction is given(pls note this is faked data. The sample size is too small to find DMCTs.):

head(celldmc.o$dmct)

The estimated coefficients for each cell-type are given in the celldmc.o$coe. Pls refer to help page of CellDMC for more info.

Sessioninfo

sessionInfo()

References

sjczheng/EpiDISH documentation built on June 24, 2024, 6:38 p.m.