TP53_co-occurrence.Rnw

<<setup_tp53, include=FALSE>>=
io = import('io')
ar = import('array')
df = import('data_frame')
st = import('stats')
tcga = import('data/tcga')
util = import('./util_2')

# define the genes we're interested in
genes_mut = c("TP53")
genes_cna = c("TP53_del", "TP53_amp", "MYC_amp", "RAD21_amp", "NDRG1_amp", "PABPC1_amp", "GATA3_amp", "SOX17_amp")

# load tcga scores for p53
p53 = io$load('../scores/tcga/pathways_mapped/speed_matrix.RData')[,'p53',drop=FALSE]

# load mutation status for the genes
path = '../analyses/tcga_mutation'
mut = io$load(file.path(path, 'mut_driver_matrix.RData'))[,genes_mut,drop=FALSE]
cna = io$load(file.path(path, 'cna_driver_matrix.RData'))[,genes_cna,drop=FALSE]
data = ar$stack(list(mut, cna), along=2)

TP53 = data[,"TP53"]
TP53_del = data[,"TP53_del"]
TP53_amp = data[,"TP53_amp"]
others = data[,!grepl("TP53", colnames(data))]

ar$intersect(p53, TP53, TP53_del, TP53_amp, data, others, along=1)
study = tcga$barcode2study(rownames(others))

# all TP53 mods are required as covariates to reach
# p>0.4 for MYC, RAD21, NDRG1, and PABPC1
assocs = st$lm(p53 ~ study + TP53 + TP53_del + TP53_amp + others) %>%
	filter(term == "others") %>%
	select(-term, -p53) %>%
	mutate(adj.p = p.adjust(p.value, method="fdr"))
@