Hi, I am using Rsubread to map Illumina RNA-seq reads, to investigate alternative splicing events. Using: 1000 Synthetic reads (size of reads are 210 bp) and a reference chromosome concatenated from all the reads and 7000 more.
My aim is to use subjunc and learn about the exon-exon junctions in BED files. I have noticed the percent of mapped reads in the align function is much higher than in the subjunc function (align is 70%, subjunc is 40%).
Why does subjunc map less reads than the align function? As I read on the manual, the align is much more flexible, but I still can't understand what is there in the subjunc function that doesn't map as much as the align function, and how can increase the percentage?
I tried:
- Playing around with the parameters (indels, mismatch, and nbestlocations)
- Merging paired end reads (using PEAR)
Unfortunately the percentage of mapped reads is still the same.
Any suggestions?
align.stat <- align(index="/Users/files_intron_rent_chrs/INDEX/intron_retention_reference_chromosomes",
readfile1=mergedreads,
output_format = "SAM",
output_file = paste("alignedmerged", "SAM", sep="."),
phredOffset = 33,
indels = 1,
maxMismatches = 1,
detectSV = TRUE,
keepReadOrder = FALSE,
minFragLength = 20,
maxFragLength = 500,
unique = FALSE,
nBestLocations = 10)
## Align function Results ##
# Total reads : 932
# Mapped : 686 (73.6%)
# Uniquely mapped : 676
# Multi-mapping : 10
# Unmapped : 246
# Fusions : 128
# Indels : 166
----------
ex_junctions <- subjunc(index ="/Users/files_intron_rent_chrs/INDEX/intron_retention_reference_chromosomes",
readfile1=mergedreads,
output_format = "SAM",
output_file = paste("alignedmerged", "SAM", sep="."),
phredOffset = 33,
indels = 1,
maxMismatches = 1,
reportAllJunctions = TRUE,
minFragLength = 20,
maxFragLength = 700,
nBestLocations = 16)
## subjunc function Results ##
# Total reads : 932
# Mapped : 381 (40.9%)
# Uniquely mapped : 374
# Multi-mapping : 7
# Unmapped : 551
# Junctions : 131
# Fusions : 0
# Indels : 89
sessionInfo( )
Thanks in advance.
Did you use the latest version of Rsubread? Can you also try to use your original paired-end reads (not merged)?