flip: Flip bins and sequences

In thackl/gggenomes: A Grammar of Graphics for Comparative Genomics

Flip bins and sequences

Description

flip and flip_seqs reverse-complement specified bins or individual sequences and their features. sync automatically flips bins using a heuristic that maximizes the amount of forward strand links between neighboring bins.

Usage

flip(x, ..., .bin_track = seqs)

flip_seqs(x, ..., .bins = everything(), .seq_track = seqs, .bin_track = seqs)

sync(x, link_track = 1, min_support = 0)

Arguments

x	a gggenomes object
...	bins or sequences to flip in dplyr::select like syntax (numeric position or unquoted expressions)
.bin_track, .seq_track	when using a function as selector such as tidyselect::where(), this specifies the track in which context the function is evaluated.
.bins	preselection of bins with sequences to flip. Useful if selecting by numeric position. It sets the context for selection, for example the 11th sequences of the total set might more easily described as the 2nd sequences of the 3rd bin: flip_seqs(2, .bins=3).
link_track	the link track to use for flipping bins nicely
min_support	only flip a bin if at least this many more nucleotides support an inversion over the given orientation

Examples

library(patchwork)
p <- gggenomes(genes=emale_genes) +
  geom_seq(aes(color=strand), arrow=TRUE) +
  geom_link(aes(fill=strand)) +
  expand_limits(color=c("-")) +
  labs(caption="not flipped")

# nothing flipped
p0 <- p %>% add_links(emale_ava)

# flip manually
p1 <- p %>% add_links(emale_ava) %>%
  flip(4:6) + labs(caption="manually")

# flip automatically based on genome-genome links
p2 <- p %>% add_links(emale_ava) %>%
  sync() + labs(caption="genome alignments")

# flip automatically based on protein-protein links
p3 <- p %>% add_sublinks(emale_prot_ava) %>%
  sync() + labs(caption="protein alignments")

# flip automatically based on genes linked implicitly by belonging
# to the same clusters of orthologs (or any grouping of your choice)
p4 <- p %>% add_clusters(emale_cogs) %>%
  sync() + labs(caption="shared orthologs")

p0 + p1 + p2 + p3 + p4 + plot_layout(nrow=1, guides="collect")

# flip seqs inside bins
s0 <- tibble::tibble(
  bin_id = c("A", "B", "B", "B", "C", "C", "C"),
  seq_id = c("a1","b1","b2","b3","c1","c2","c3"),
  length = c(1e4, 6e3, 2e3, 1e3, 3e3, 3e3, 3e3))

p <- gggenomes(seqs=s0) +
  geom_seq(aes(color=bin_id), size=1, arrow = arrow(angle = 30, length = unit(10, "pt"),
    ends = "last", type = "open")) +
  geom_bin_label() + geom_seq_label() +
  expand_limits(color=c("A","B","C"))

p1 <- p %>% flip_seqs(6)
p2 <- p %>% flip_seqs(c2)
p3 <- p %>% flip_seqs(2, .bins = C)

p + p1 + p2 + p3 + plot_layout(nrow=1, guides="collect")

# fancy flipping using tidyselect::where for dynamic selection
p <- gggenomes(emale_genes,emale_seqs) %>% add_clusters(emale_cogs) +
  geom_seq(color="grey70", size=1, arrow = arrow(angle = 30, length = unit(15, "pt"),
    ends = "last", type = "open")) +
  geom_gene(aes(fill=cluster_id))

# flip all short seqs - where() applied to .bin_track=seqs
p1 <- p %>% flip(where(~.x$length < 21000))

# flip all seqs with MCP on "-" - where() applied to .bin_track=genes
p2 <- p %>% flip(where(~any(.x$strand[.x$cluster_id %in% "cog-MCP"] == "-")), .bin_track=genes)

p + p1 + p2 + plot_layout(nrow=1, guides="collect") & theme(legend.position = "bottom")

thackl/gggenomes documentation built on March 10, 2024, 7:26 a.m.