connect_satalite_vertices: Connect integration sites which lie a specific gap distance...

In cnobles/gintools: Genomic DNA Integration Analysis Tools

Description

Given a list of unique integration site positions (reduced GRanges object) and a directed graph of connected components, this function identifies each node within gap range of clusters and creates an edge to connect the cluster to the 'satalite' node. Edges are drawn from the last node in the cluster to the 'satalite' node, but directionality is determined first by abundance and secondly by an upstream bias.

Usage

connect_satalite_vertices(red.sites, graph, gap, bais)

Arguments

red.sites	GRanges object which has been reduced to single nt positions and contains the revmap from the original GRanges object. The object must also contain a column for cluster membership (clus.id) and a column for abundance (fragLengths).
graph	a directed graph built from the red.sites object. Each node corresponds to a row in the red.sites object.
gap	integer nucleotide (nt) gap distance to consider joining to clusters.
bias	either "upsteam" or "downstream", designating which position to choose if other decision metrics are tied.

Details

connect_satalite_vertices returns a graph where nodes within 'gap' distance from clusters are now connected to the edge of the clusters.

Author(s)

Christopher Nobles, Ph.D.

Examples

gr <- gintools:::generate_test_granges(stdev = 3)
red.sites <- reduce(
  flank(gr, -1, start = TRUE),
  min.gapwidth = 0L,
  with.revmap = TRUE)
red.sites$siteID <- seq_along(red.sites)
revmap <- as.list(red.sites$revmap)
red.sites$fragLengths <- lengths(revmap)
red.hits <- GenomicRanges::as.data.frame(
  findOverlaps(red.sites, maxgap = 0L, drop.self = TRUE))
red.hits <- red.hits %>%
  mutate(q_pos = start(red.sites[queryHits])) %>%
  mutate(s_pos = start(red.sites[subjectHits])) %>%
  mutate(q_fragLengths = red.sites[queryHits]$fragLengths) %>%
  mutate(s_fragLengths = red.sites[subjectHits]$fragLengths) %>%
  mutate(strand = unique(strand(
    c(red.sites[queryHits], red.sites[subjectHits])))) %>%
  mutate(is.upstream = ifelse(
    strand == "+",
    q_pos < s_pos,
    q_pos > s_pos)) %>%
  mutate(keep = q_fragLengths > s_fragLengths) %>%
  mutate(keep = ifelse(
    q_fragLengths == s_fragLengths,
    is.upstream,
    keep)) %>%
  filter(keep)
g <- make_empty_graph(n = length(red.sites), directed = TRUE) %>%
  add_edges(unlist(mapply(
    c, red.hits$queryHits, red.hits$subjectHits, SIMPLIFY = FALSE)))
red.sites$clus.id <- clusters(g)$membership

connect_satalite_vertices(red.sites, g, gap = 2L, "upstream")

cnobles/gintools documentation built on Aug. 22, 2019, 10:36 a.m.