In BBCG/epialleleR: Fast, Epiallele-Aware Methylation Caller and Reporter

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  width = 100
)
options(width=100)

Introduction

The best possible explanation on VEF and lMHL values is given in help files for generateCytosineReport and generateMhlReport methods, respectively. Here we try to show some simplified and real situations, i.e., different methylation patterns that may exist, and provide a visual summary of epialleleR output.

The readers are welcome to try their own real and simulated data. If it might be of interest to others, please create an issue and these examples might get included in this vignette.

NB: the plotMetrics function used below is a piece of spaghetti code, hence hidden. If you still want to use it or see what it does - browse a source code of this vignette online.

require("data.table", quietly=TRUE)
require("GenomicRanges", quietly=TRUE)
require("ggplot2", quietly=TRUE)
require("gridExtra", quietly=TRUE)
require("epialleleR", quietly=TRUE)

plotMetrics <- function (bam.file, range, min.n=0, title="epialleles") {
  bam <- preprocessBam(bam.file=bam.file, verbose=FALSE)
  cg.beta <- generateCytosineReport(bam, threshold.reads=FALSE, verbose=FALSE)
  cg.vef  <- generateCytosineReport(bam, threshold.reads=TRUE, verbose=FALSE)
  cg.mhl  <- generateMhlReport(bam, max.haplotype.window=20, verbose=FALSE)
  range.strand <- as.character(strand(range))
  if (range.strand=="*") range.strand <- c("+", "-")
  metrics <- cbind(
    cg.beta[rname==as.character(seqnames(range)) & pos>=start(range) & pos<=end(range) & strand %in% range.strand,
            .(pos=factor(pos), beta=meth/(meth+unmeth))],
    cg.vef[rname==as.character(seqnames(range)) & pos>=start(range) & pos<=end(range)  & strand %in% range.strand,
           .(VEF=meth/(meth+unmeth))],
    cg.mhl[rname==as.character(seqnames(range)) & pos>=start(range) & pos<=end(range)  & strand %in% range.strand,
           .(lMHL=lmhl)]
  )

  metrics.melt <- melt.data.table(metrics, id.vars="pos")
  metrics.melt[value<0.00001, value:=0]
  # metrics.melt[, class:=ifelse(variable %like% "gMHL", 1, 0.5)]

  patterns <- extractPatterns(bam, range, verbose=FALSE)[strand %in% range.strand]

  base.positions <- grep("^[0-9]+$", colnames(patterns), value=TRUE)
  patterns.summary <- patterns[, c(lapply(.SD, unique), .N),
                               by=.(pattern, beta), .SDcols=base.positions]
  plot.data <- data.table::melt.data.table(patterns.summary,
                                           measure.vars=base.positions, variable.name="pos", value.name="base")
  plot.data <- na.omit(plot.data)[N>=min.n]

  base.positions <- as.numeric(as.character(base.positions))
  plot.data[, pos:=as.numeric(as.character(pos))]
  metrics.melt[, pos:=as.numeric(as.character(pos))]

  # upset-like plot of all patterns, categorical positions, sorted by counts
  if (require("ggplot2", quietly=TRUE) & require("gridExtra", quietly=TRUE)){
    epialleles <- ggplot(plot.data,
                         aes(x=pos, y=reorder(pattern,N),
                             color=factor(base, levels=c("z","Z")))) +
      geom_line(color="grey") +
      geom_point() +
      scale_colour_grey(start=0.8, end=0) +
      theme_light() +
      # scale_x_continuous(breaks=base.positions[c(1, length(base.positions))]) +
      theme(axis.text.y=element_blank(), legend.position="none") +
      labs(x="position", y=NULL, title=title, color="base")

    bars <- ggplot(unique(plot.data[, .(pattern, N, beta)]),
                   aes(x=N+0.5, y=reorder(pattern,N), alpha=beta, label=N)) +
      geom_col(alpha=pmin(unique(plot.data[, .(pattern, N, beta)])$beta+0.1, 1)) +
      geom_text(alpha=0.5, hjust=0, nudge_x=0.25, size=3) +
      scale_x_log10(expand = expansion(mult=c(0, 0.3))) +
      theme_minimal() +
      theme(axis.text.y=element_blank(), legend.position="none") +
      labs(x="count", y=NULL, title="")

    values <- ggplot(metrics.melt,
                     aes(x=pos, y=value, color=variable, group=variable)) +
      geom_line(alpha=0.5, linewidth=1) + #metrics.melt$class) +
      scale_color_brewer(palette="Set1") +
      # scale_x_discrete(breaks=base.positions[c(1, length(base.positions))]) +
      scale_y_continuous(trans="log10", limits=c(0.00001,1), breaks=10**-c(0:5)) +
      theme_light() +
      theme(legend.position="bottom")

    epi.grob <- ggplotGrob(epialleles)
    bar.grob <- ggplotGrob(bars)
    val.grob <- ggplotGrob(values)
    max.widths <- do.call(grid::unit.pmax, lapply(list(epi.grob, val.grob), `[[`, "widths"))
    epi.grob$widths[2:5] <- max.widths[2:5]
    grid.arrange(
      epi.grob, bar.grob, val.grob, NULL,
      ncol=2, widths=c(0.85, 0.15), heights=c(3, 2)
    )
  }
}

out.bam <- tempfile(pattern="simulated", fileext=".bam")

# no epimutations
simulateBam(
  output.bam.file=out.bam,
  XM=c(
    sapply(
      lapply(1:1000, function (x) sample(c("Z",rep("z", 9)), 10)),
      paste, collapse=""
    )
  ),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-10", "GRanges"), 0, title="no epimutations")

# one complete epimutation
simulateBam(
  output.bam.file=out.bam,
  XM=c(
    paste(rep("Z", 10), collapse=""),
    sapply(
      lapply(1:999, function (x) sample(c("Z",rep("z", 9)), 10)),
      paste, collapse=""
    )
  ),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-10", "GRanges"), title="one complete epimutation")

# one partial epimutation
simulateBam(
  output.bam.file=out.bam,
  XM=c(
    paste(c(rep("Z", 4), "z", "z", rep("Z", 4)), collapse=""),
    sapply(
      lapply(1:999, function (x) sample(c("Z",rep("z", 9)), 10)),
      paste, collapse=""
    )
  ),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-10", "GRanges"), title="one partial epimutation")

# another partial epimutation
simulateBam(
  output.bam.file=out.bam,
  XM=c(
    "zZZZZZZZzz",
    sapply(
      lapply(1:999, function (x) sample(c("Z",rep("z", 9)), 10)),
      paste, collapse=""
    )
  ),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-10", "GRanges"), title="another partial epimutation")

# several partial epimutations
simulateBam(
  output.bam.file=out.bam,
  XM=c(
    sapply(
      lapply(1:10, function (x) c(rep("Z", 6), rep("z", 4))),
      paste, collapse=""
    ),
    sapply(
      lapply(1:999, function (x) sample(c("Z",rep("z", 9)), 10)),
      paste, collapse=""
    )
  ),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-10", "GRanges"), title="several partial epimutations")

# several short partial epimutations
simulateBam(
  output.bam.file=out.bam,
  XM=c(
    sapply(
      lapply(1:10, function (x) c(rep("Z", 4), rep("z", 6))),
      paste, collapse=""
    ),
    sapply(
      lapply(1:999, function (x) sample(c("Z",rep("z", 9)), 10)),
      paste, collapse=""
    )
  ),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-10", "GRanges"), title="several short partial epimutations")

# several overlapping partial epimutations
simulateBam(
  output.bam.file=out.bam,
  pos=1:10,
  XM=c(
    "ZZZZZZZZZZ", "ZZZZZZZZZz", "ZZZZZZZZzz", "ZZZZZZZzzz", "ZZZZZZzzzz",
    sapply(
      lapply(1:15, function (x) sample(c("Z",rep("z", 9)), 10)),
      paste, collapse=""
    )
  ),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-20", "GRanges"), title="several overlapping partial epimutations")

# amplicon 0%
plotMetrics(
  system.file("extdata", "amplicon000meth.bam", package="epialleleR"),
  as("chr17:43124861-43126026", "GRanges"), title="amplicon, 0%"
)

# amplicon 10%
plotMetrics(
  system.file("extdata", "amplicon010meth.bam", package="epialleleR"),
  as("chr17:43124861-43126026", "GRanges"), title="amplicon, 10%"
)

# sample capture, BMP7
plotMetrics(
  system.file("extdata", "capture.bam", package="epialleleR"),
  as("chr20:57266125-57268185:+", "GRanges"), title="sample capture, BMP7, + strand"
)

# sample capture, BMP7
plotMetrics(
  system.file("extdata", "capture.bam", package="epialleleR"),
  as("chr20:57266125-57268185:-", "GRanges"), title="sample capture, BMP7, - strand"
)

# sample capture, RAD51C
plotMetrics(
  system.file("extdata", "capture.bam", package="epialleleR"),
  as("chr17:58691673-58693108:+", "GRanges"), title="sample capture, RAD51C, + strand"
)

# sample capture, RAD51C
plotMetrics(
  system.file("extdata", "capture.bam", package="epialleleR"),
  as("chr17:58691673-58693108:-", "GRanges"), title="sample capture, RAD51C, - strand"
)

# long-read sequencing, low methylation
getXM <- function (p) {sample(x=c("z", "Z"), size=1, prob=c(p, 1-p))}
probs <- (sin(seq(-2*pi, +1*pi, by = pi/125))+2)/3
simulateBam(
  output.bam.file=out.bam,
  pos=1:10,
  XM=sapply(1:10, function (i) {paste(sapply(probs, getXM), collapse="")}),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-1000", "GRanges"), title="long-read sequencing, low methylation")

# long-read sequencing, high methylation
simulateBam(
  output.bam.file=out.bam,
  pos=1:10,
  XM=sapply(1:10, function (i) {paste(sapply(1-probs, getXM), collapse="")}),
  XG="CT"
)
plotMetrics(out.bam, as("chrS:1-1000", "GRanges"), title="long-read sequencing, high methylation")