In this tutorial there's a part in which they try to shuffle the lines of a file and then split the file into two. I do not understand why when splitting the file into 2, instead of dividing the number of lines by two they add 1 and then divide by two. Why is that?
That part of the code is:
nlines=$(samtools view merged.bam | wc -l ) # Number of reads in the BAM file
nlines=$(( (nlines + 1) / 2 )) # half that number
samtools view ${tmpDir}/${NAME1}_${NAME2}_merged.bam | shuf - | split -d -l ${nlines} - "${tmpDir}/${EXPT}" # This will shuffle the lines in the file and split it
into two SAM files
cat ${tmpDir}/${EXPT}_header.sam ${tmpDir}/${EXPT}00 | samtools view -bS - > ${outputDir}/${EXPT}00.bam
cat ${tmpDir}/${EXPT}_header.sam ${tmpDir}/${EXPT}01 | samtools view -bS - > ${outputDir}/${EXPT}01.bam
1 answer
Think I know now:
In bash it only gives the integer part of the number. If the original file has for e.g. 3 lines and we want to split in half, we want one file to have 2 lines and the other 1 line, but 3/2=1.5 and bash considers to be =1, so then the file will be split in 3 files each one with one line. If we add 1 before dividing by 2: (3+1)/2=2, so bash will put two lines into a file and the remaining 1 line in the other, as we wanted.
If the file has an even number of lines: for e.g original file has 4 lines, (4+1)/2=2.5 and bash thinks it's 2, so file will be split into files which will contain 2 lines each, so still does what we want for even numbers.
Log in to answer this question.
-
Pseudoreplicate creation from bam files
written by sarahmanderni
-
paired-end reads, idxstats, and samtools and grep
written by Bogdan
-
Differing number of reads for read 1 and 2 in fastq's from a subset bam file
written by michael •
-
Splitting human bam file into fixed regions
written by irfanwustl •
-
Header error while converting sam file to bam
written by MS •
-
Chimeric read insert length in SAM $9
written by Yijun Tian •
-
Piping bowtie2 output directly into BAM
written by hersh •
-
Creating pseudo-replicates from a paired-end BAM file
written by James Ashmore
-
Split SAM file into smaller SAM files by line count
written by cacampbell •
-
recommended way of filtering out non-karyotype ref sequences and reads from a bam file?
written by 14134125465346445
Are there an odd number of lines in the bam?
You can try for yourself, run the relevant part with part of the tutorial with
nlines=$(( (nlines + 1) / 2 ))andnlines=$(( (nlines ) / 2 )), and see if it makes a difference.You may also open an issue at the repo with your question. But please first try for yourself and check for differences, if any.