R("""
library(ggplot2)
library(scales)
library(gridExtra)
data = tryCatch({{return(read.table("%(infile)s", header=F))}}, error=function(e) {{return(data.frame())}})
if (dim(data)[1] == 0) {{
pdf("%(outfile)s")
plot(0:10, type = "n", xaxt="n", yaxt="n", bty="n", xlab = "", ylab = "")
text(5, 8, "No data", cex=2)
dev.off()
}} else {{
colnames(data) = c("seqId", "len")
plotList = list()
plotList[[1]] = ggplot(data, aes(x=len)) + geom_histogram(binwidth=1, color="black") + ggtitle(expression(atop("Length distribution of predicted amplicons", atop("with primer pair: %(primers)s on database: %(db)s", atop("min. Temp.: %(minT).2f, max. Temp.: %(maxT).2f, max. Len.: %(len)i"))))) + xlab("length")
plotList[[2]] = ggplot(data, aes(x=len)) + geom_density() + ggtitle(expression(atop("Length distribution of predicted amplicons", atop("with primer pair: %(primers)s on database: %(db)s", atop("min. Temp.: %(minT).2f, max. Temp.: %(maxT).2f, max. Len.: %(len)i"))))) + xlab("length")
pdf("%(outfile)s")
print(plotList)
dev.off()
}}
""" % {"infile": input[0], "outfile": output[0],
"primers": wildcards.pair_name,
"db": wildcards.dbname,
"len": config["primer_pairs"][wildcards.pair_name]["max_len"],
"maxT":config["primer_pairs"][wildcards.pair_name]["max_temp"],
"minT": config["primer_pairs"][wildcards.pair_name]["min_temp"],
}
)
R("""
library(SKAT)
library(stringr)
library(plotrix)
assoc <- read.delim('{input.assoc}', sep=" ")
n <- min(nrow(assoc),16)
if (n > 0) {{
for (i in 1:n) {{
row <- assoc[i, ]
ssd = paste('ssd/{wildcards.dataset}.chr',row$CHR,'.SSD', sep="")
fam = paste('bed/{wildcards.dataset}.chr',row$CHR,'.fam', sep="")
info = paste('ssd/{wildcards.dataset}.chr',row$CHR,'.Info', sep="")
pheno <- '{input.pheno}'
FAM<- Read_Plink_FAM_Cov(fam, pheno, Is.binary=FALSE)
y <- FAM${wildcards.pheno}
SSD.INFO = Open_SSD(File.SSD=ssd, File.Info=info)
SetIndex <- which(SSD.INFO$SetInfo$SetID == row$SetID)
genos <- Get_Genotypes_SSD(SSD.INFO, SetIndex, TRUE)
genos <- as.data.frame(genos)
genos${wildcards.pheno} <- y
genos$haplotypes <- do.call(paste0, genos[colnames(genos)[1:length(colnames(genos))-1]])
genos$haplotypes <- as.factor(genos$haplotypes)
png(paste('{output}',str_pad(i, 2, pad="0"),'_',row$SetID,'.png',sep=""))
boxplot(genos${wildcards.pheno} ~ genos$haplotypes, main=paste('{wildcards.dataset}',' chr', row$CHR, sep=""), ylab='{wildcards.pheno}', xlab=row$SetID)
colors <- ifelse(FAM$braak > 3, 'red', 'green')
points(genos${wildcards.pheno} ~ genos$haplotypes, col=colors)
mtext(paste(colnames(genos)[1:max(1, ncol(genos)-2)], collapse=" "),side=3,outer=F)
# count genotypes
f <- summary(genos$haplotypes)
f <- as.data.frame(f)
f <- cbind(rownames(f),f$f)
colnames(f) <- c('Hap','n')
f <- apply(f, 2, str_pad, width=nchar(f[1,1]), pad=" ")
addtable2plot(x="bottomright",table=f)
dev.off()
}}
}}
""")
workdir: "workflow4"
from snakemake.utils import R
SAMPLES = ["GSM521934", "GSM521935"]
rule all:
input: expand("{sample}_sorted.bam", sample = SAMPLES)
rule sam_to_bam:
input: "{file}.sam"
output: "{file}.bam"
params: threads = 2
log: "{file}.log"
benchmark: "{file}.json"
shell: "(samtools view -bS --threads {params.threads} {input} > {output}) > {log}"
rule bam_sorted:
input: "{file}.bam"
output: "{file}_sorted.bam"
run:
R("""
library(Rsamtools)
library(tools)
sortBam("{input}", "{output}")
file.rename("{output}.bam", file_path_sans_ext("{output}.bam"))
""")
def boxplot(input, output, x, y, x_log_scale_base=None, y_log_scale_base=None, remove_na=False, remove_outlier=False):
command = """
library(ggplot2)
library(scales)
d <- read.delim("{input}", header=T)
"""
if remove_na:
command += """
d <- d[complete.cases(d),]
"""
command += """
g <- ggplot(d, aes(x={x}, y={y}, group={x}))
"""
if remove_outlier:
command += """
g <- g + geom_boxplot(outlier.shape = NA)
ylim1 <- boxplot.stats(d${y})$stats[c(1, 5)]
# g <- g + geom_point()
"""
else:
command += """
g <- g + geom_boxplot()
# g <- g + geom_point()
"""
# if ((x_log_scale_base=="10") and (y_log_scale_base=="10")):
# command += """
# g <- g + coord_trans(x = "log10", y = "log10")
# """
if x_log_scale_base is not None:
command += """
g <- g + scale_x_continuous(
trans = 'log{x_log_scale_base}',
breaks = pretty_breaks()
# labels = trans_format('log{x_log_scale_base}', math_format({x_log_scale_base}^.x))
)
"""
if y_log_scale_base is not None:
command += """
g <- g + scale_y_continuous(
trans = 'log{y_log_scale_base}',
breaks = pretty_breaks()
# labels = trans_format('log{y_log_scale_base}', math_format({y_log_scale_base}^.x))
)
"""
if remove_outlier:
command += """
g <- g + coord_cartesian(ylim = ylim1*1.05)
"""
# command += """
# ggsave(file="{output}", plot=g)
# """
command += """
pwd <- getwd()
setwd(dirname("{output}"))
ggsave(file=basename("{output}"), plot=g)
setwd(pwd)
"""
command = command.format(input=input, output=output, x=x, y=y, x_log_scale_base=x_log_scale_base, y_log_scale_base=y_log_scale_base)
print(command)
R(command)
def summmarized_plot(input, output, x, y1, y2, y1_op="sd", y2_op="mean", x_log_scale_base=None, y_log_scale_base=None, remove_na=False):
command = """
library(ggplot2)
library(scales)
library(plyr)
d <- read.delim("{input}", header=T)
"""
if remove_na:
command += """
d <- d[complete.cases(d),]
"""
command += """
d_stat <- ddply(d, .({x}), summarize, "{y1_op}_{y1}"={y1_op}({y1}), "{y2_op}_{y2}"={y2_op}({y2}))
d1 <- data.frame(d_stat${x}, d_stat${y1_op}_{y1}, "{y1_op}_{y1}")
colnames(d1) <- c("{x}", "{y1}", "group")
d2 <- data.frame(d_stat${x}, d_stat${y2_op}_{y2}, "{y2_op}_{y2}")
colnames(d2) <- c("{x}", "{y1}", "group")
d_merged <- rbind(d1, d2)
g <- ggplot(d_merged, aes(x={x}, y={y1}, group=group, color=group))
g <- g + geom_point()
g <- g + geom_line()
"""
# if ((x_log_scale_base=="10") and (y_log_scale_base=="10")):
# command += """
# g <- g + coord_trans(x = "log10", y = "log10")
# """
if x_log_scale_base is not None:
command += """
g <- g + scale_x_continuous(
trans = 'log{x_log_scale_base}',
breaks = pretty_breaks()
# breaks = trans_breaks('log{x_log_scale_base}', function(x) {x_log_scale_base}^(x/2)),
# labels = trans_format('log{x_log_scale_base}', math_format({x_log_scale_base}^.x))
)
"""
if y_log_scale_base is not None:
command += """
g <- g + scale_y_continuous(
trans = 'log{y_log_scale_base}',
breaks = pretty_breaks()
# breaks = trans_breaks('log{y_log_scale_base}', function(x) {y_log_scale_base}^(x/2)),
# labels = trans_format('log{y_log_scale_base}', math_format({y_log_scale_base}^.x))
)
"""
# command += """
# g <- g + ylab(NULL)
# g <- g + scale_color_hue(name="", labels=c({y1_op}_{y1}="{y1_op}({y1})", {y2_op}_{y2}="{y2_op}({y2})"))
# ggsave(file="{output}", plot=g)
# """
command += """
g <- g + ylab(NULL)
g <- g + scale_color_hue(name="", labels=c({y1_op}_{y1}="{y1_op}({y1})", {y2_op}_{y2}="{y2_op}({y2})"))
"""
command += """
pwd <- getwd()
setwd(dirname("{output}"))
ggsave(file=basename("{output}"), plot=g)
setwd(pwd)
"""
command = command.format(input=input, output=output, x=x, y1=y1, y2=y2, y1_op=y1_op, y2_op=y2_op, x_log_scale_base=x_log_scale_base, y_log_scale_base=y_log_scale_base)
print(command)
R(command)
R("""
library('reshape')
library('ggplot2')
library('edgeR')
setwd("{DIR}")
myfiles=as.character(unlist(strsplit("{input.files}", split=" ")))
res=read.delim(myfiles[1],header=T)
colnames(res)[1]="gene"
colnames(res)[2]=as.character(myfiles[1])
# remove the last 5 statistics lines ...
# nr=dim(res)[1]
# res=res[-c((nr-4):nr),]
#
for(i in seq(2, length(myfiles), by = 1))
{{
temp=read.delim(myfiles[i],header=T)
colnames(temp)[1]="gene"
colnames(temp)[2]=as.character(myfiles[i])
res=merge(res,temp)
}}
write.table(as.data.frame(res),file="RawCountFile.txt",sep="\t",row.names=F)
#
mydata=read.delim("RawCountFile.txt",row.names=1)
val1=as.numeric("{MINCOUNT}")
val2=as.numeric("{MINSAMPLES}")
cat(val1," ", val2, "checking..\n",file="check.txt")
filter <- apply(mydata, 1, function(x) length(x[x>val1])>=val2)
res=mydata[filter,]
write.table(as.data.frame(res),file="RawCountFile_filtered.txt",sep="\t",col.names=NA)
png("HistBeforenormFilter.png")
df.m <- melt(as.data.frame(res))
print(ggplot(df.m) + geom_density(aes(x = value, colour = variable)) + labs(x = NULL) + theme(legend.position='top') + scale_x_log10())
dev.off()
y = DGEList(counts=res)
## Normalization TMM ------------------------------------------------------------
## method = =c("TMM","RLE","upperquartile","none")
y <- calcNormFactors(y,method="TMM")
ndata= cpm(y,log=FALSE,normalized.lib.sizes=TRUE)
## save it
write.table(ndata,file="CPM_TMM_counts.txt",sep="\t",col.names=NA)
""")
out.write("\t".join([",".join(seqList) for seqList in listTab.values()]))
with open(output.counts, "w") as out:
for name, seqList in listTab.items():
out.write("%s\t%i\n" % (name, len(seqList)))
with open(output.readOtuMap, "w") as out:
for otu, seqList in listTab.items():
for seq in seqList:
out.write("%s\t%s\n" % (seq, otu))
rule plotOtuSizeDist:
input: "{sample}.full.good.unique.abund.otus.counts.tsv"
output: "{sample}.full.good.unique.abund.otus.sizeDist.pdf"
run:
R("library(ggplot2)\nd=read.table(\"%s\")\npd = as.data.frame(cbind(seq(1,length(d$V2)), d$V2[order(d$V2, decreasing=T)]))\np=ggplot(pd) + geom_segment(aes(x=V1, xend=V1, y=V2, yend=0)) + xlab(\"OTU rank\") + ylab(\"size\") + ggtitle(\"%s - OTU size distribution\")\nggsave(\"%s\", p)" % (input[0], "{wildcards.sample}", output[0]))
rule computRarefaction:
input: "{sample}.full.good.unique.abund.otus.list"
output: "{sample}.full.good.unique.abund.otus.rarefaction"
shell:
"%(mothur)s \"#rarefaction.single(list={input})\"" % config
rule plotRarefaction:
input: "{sample}.full.good.unique.abund.otus.rarefaction"
output: "{sample}.full.good.unique.abund.otus.rarefaction.pdf"
run:
R("""library(ggplot2)
d=read.table("%s", header=T)
p = ggplot(d, aes(numsampled)) + geom_point(aes(y=X0.03), colour="blue", size=1) + geom_ribbon(aes(ymin=lci, ymax=hci), alpha=0.2) + xlab("number of reads sampled") + ylab("number of OTUs observed") + ggtitle("Rarefaction curve - %s")
rule quantification_with_featureCounts:
input: novel="samples/new_annotation/all_transcripts.gtf", bam=expand("samples/bam/{smp}.bam", smp=SAMPLES)
output: "results/counts/gene_counts.txt", "results/counts/gene_counts_mini.txt"
shell: """
featureCounts -p -s 2 -T 15 -t exon -g gene_id -a {input.novel} -o {output[0]} {input.bam} &> {output[0]}.log
cut -f 1,7- {output[0]}| awk 'NR > 1' | awk '{{gsub("samples/bam/","",$0); print}}' > {output[1]}
"""
rule diagnostic_plot:
input: "results/counts/gene_counts_mini.txt"
output: "results/diagnostic_plot/diagnostic.pdf"
run: R("""
dir.create("results/diagnostic_plot")
data <- read.table("{input}",
sep="\t",
header=T,
row.names=1)
data <- data[rowSums(data) > 0, ]
data <- log2(data + 1)
pdf("{output}")
dev.null <- apply(data, 2, hist, border="white", col="blue")
boxplot(data, color="blue", pch=16)
pairs(data, pch=".", col="blue")
dev.off()
cat("etc...")
""")
rule limit_for_plot:
input: WORK + '{method}.eval_{dat}.{cols}.eval'
output: WORK + '{method}.eval_{dat}.{cols}.totWrong'
shell: "grep 'TotWrong\|count' {input} | grep -v ssue | grep -v earing > {output}"
rule cat:
input: expand( WORK + '{{method}}.eval_{dat}.{{cols}}.totWrong', dat=('clinvar', 'denovo', 'clinvar_mult', 'clinvar_single', 'clinvar_exp') )
output: o = WORK + 'totWrong/{method}.{cols}'
run:
pd.concat( [pd.read_csv(x, sep='\t') for x in list(input)] ).to_csv(output.o, index=False, sep='\t')
rule plot:
input: WORK + 'totWrong/{method}.{cols}'
output: DOCS + 'plot/{method}.other.{cols}.totWrong.png'
run:
R("""
require(ggplot2)
d = read.delim("{input}", sep='\t', header=TRUE)
p = ggplot(data=d) +
geom_col(aes(y=var_count,x=score_type, fill=score_type)) +
facet_grid(clinvar_type~., scale='free') + theme_bw() +
ylab('Wrong Predictions') +
theme(axis.text.x = element_text(angle=45, hjust=1)) +
xlab('') + theme(legend.position="none")
ggsave("{output}", p)
""")
rule all_eval:
input: expand( DOCS + 'plot/{method}.other.{cols}.totWrong.png', method=('global',), cols=('mpc', 'revel', 'mpc-revel', 'ccr', 'mpc-revel-ccr', 'mpc-ccr', 'revel-ccr') )
out.write("%s\t%i\tssu\t%s\t%i\n" % (oId, size, "\t".join([ssuCls[r] for r in ranks]), ssuDepth))
out.write("%s\t%i\t**s\t%s\t%i\n" % (oId, size, "\t".join([itsCls[r] for r in ranks]), itsDepth))
out.write("%s\t%i\tlsu\t%s\t%i\n" % (oId, size, "\t".join([lsuCls[r] for r in ranks]), lsuDepth))
rule clsSummary:
"""Collect some summary stats of how many OTUs were assigend to taxonomic
ranks for the paper abstract"""
input: "taxonomy/{sampleSet}_97_comb.stats.tsv"
output: "taxonomy/{sampleSet}_97_clsStats.tsv"
run:
R("""
d=read.table("{input}", sep="\t")
colnames(d) = c("oId", "size", "mrk", "domain", "kingdom", "phylum", "class", "order", "family", "genus", "species", "depth")
a=aggregate(depth ~ oId, subset(d, size>1), max)
ranks=c("kingdom", "phylum", "class", "order", "family", "genus", "species")
s = data.frame(rank=numeric(0), nubmer=numeric(0))
for (i in 1:7) {{
s = rbind(s, data.frame(rank=ranks[i], number=sum(a$depth>=i)))
}}
write.table(s, "{output}", sep="\t", row.names=F)
""")
rule plotClsComp:
"""Create plots of classifications depth"""
input: all="taxonomy/{sampleSet}_97_comb.stats.tsv"
output: depth="{sampleSet}_clsComp_depth.svg", depthFungi="{sampleSet}_clsComp_depth_fungi.svg", block="{sampleSet}_clsComp_basic.svg"
run:
R("""
library(reshape2)
library(ggplot2)