drug_volcanos.Rnw

<<setup3, include=FALSE>>=
library(dplyr)
b = import('base')
io = import('io')
ar = import('array')
st = import('stats')
df = import('data_frame')
plt = import('plot')
gdsc = import('data/gdsc')
util = import('./util_3')
config = import('../config')

assocs = b$lnapply(config$methods$analysis_set, util$vp$load_fun)
@

\begin{figure}[H]
<<drug_volcano_pancan, echo=FALSE, fig.width=6.5, fig.height=8.5>>=
p = assocs %>%
    b$lnapply(function(x) x$pan) %>%
    df$bind_rows(name_col="method") %>%
    mutate(label = paste(drug, scores, sep=":"),
           method = config$id2short(method)) %>%
    plt$color$p_effect(pvalue="adj.p", effect="estimate", thresh=0.05, dir=-1) %>%
    plt$volcano(base.size=0.03, text.size=2, p=0.05, ceil=1e-15, label_top=130, repel=TRUE) +
        facet_wrap(~method, ncol=2, scales="free_x") +
        config$facet_theme
plt$symmetrise_scale(p)
@
\caption{Volcano plots for associations between pathway scores and drug
    response (IC50). Effect size arbitrary units, p-values FDR-corrected.
        Pathway-response genes are the only method to recover highly
        significant oncogene addiction associations, the rest of methods show
        no obvious connection between the drug target and pathway.}
\label{fig:drug-volcano-pancan}
\end{figure}


%\begin{figure}[H]
%<<drug_volcano_tissue, echo=FALSE, fig.width=6.5, fig.height=8.5>>=
%p = assocs %>%
%    b$lnapply(function(x) x$assocs.tissue) %>%
%    df$bind_rows(name_col="method") %>%
%    filter(subset == "clinical") %>%
%    mutate(label = paste(drug, scores, tissue, sep=":"),
%           method = config$id2short(method)) %>%
%    plt$color$p_effect(pvalue="adj.p", effect="estimate", thresh=0.1, dir=-1) %>%
%    plt$volcano(base.size=1, text.size=2, p=0.1, label_top=60, repel=TRUE) +
%        facet_wrap(~method, ncol=2, scales="free_x") +
%        config$facet_theme
%plt$symmetrise_scale(p)
%@
%\caption{Cancer-specific volcano plots, legend as in previous figure. Again,
%    top hits for pathway-response genes are in line with the oncogene addiction
%pardigm, other methods are not. Pathifier provides highly significant
%associations for Cetuximab and Trametinib for all pathways, calling into
%question the meaning of these associations.}
%\label{fig:drug-volcano-tissue}
%\end{figure}