Entering edit mode
4.3 years ago
aboyd003
•
0
I am analyzing ChIP data. I called peaks with MACS2 run through a RECAP wrapper script. I am attempting differential binding analysis using DiffBind. Code shown below. Please note I have tried the suggestion in biocondutor support from Stark (the developer) to explicitly set the datatype as part of dba.report.
MADSmeta <- dba(sampleSheet="SampleSheet.csv")
MADScounts <- dba.count(MADSmeta)
MADScontrast <- dba.contrast(MADScounts, categories=DBA_CONDITION, block=c(DBA_TISSUE,DBA_TREATMENT))
dba.show(MADScontrast,bContrast=T)
MADSanalyze <- dba.analyze(MADScontrast, method=DBA_DESEQ2, bSubControl=FALSE, bFullLibrarySize=TRUE, bTagwise=FALSE, bCorPlot=TRUE,bReduceObjects=FALSE)
MADSreport <- dba.report(MADSanalyze, contrast=1, method=DBA_DESEQ2, th=1, DataType=DBA_DATA_FRAME)
write.table(MADSreport, file="File.csv", col.names=TRUE, row.names=FALSE, sep="\t")
The resulting output has coordinates outside of the contig sizes. Peak in the DiffBind results shows up is
7000000037415150 5009119 5009427
But the contig size (listed below) is substantially less than the start of the peak.
7000000037415150 46831
This problem also occurs when I use diffbind under windows, but when I use diffbind under Linux, it will calculate quickly.