illumina paired-end data

[sbdk82]

I have paired-end illumina data for some RNA-Seq analysis. There are four fasta files and are named like this.

XXX_XXXXXX_L001_R1_001.fastq            XXX_XXXXXX_L001_R2_001.fastq            
XXX_XXXXXX_L002_R1_001.fastq            XXX_XXXXXX_L002_R2_001.fastq

Can anyone explain what do L001 and L002 stand for? I know R1 and R2 stand for two pairs. When I use paired end reads in some tools/packages, Do I need to use four of them? or only two?

[GenoMax]

L001 and L002 stand for the lanes on which this sample ran.

[Brian Bushnell]

You don't NEED to use all four of them (and most programs won't accept 4 inputs). But if they are from the same library, then yes, you should probably merge the files from the different lanes so that you only have 1 file for read 1 and 1 file for read 2; it will simplify things downstream.

[sbdk82]

Thanks Brian and GenoMax

[GenoMax]

Merge them in the same order to keep odd things from happening later on. If you trim, use programs that will keep the mates in sync.

$ cat XXX_XXXXXX_L001_R1_001.fastq XXX_XXXXXX_L002_R1_001.fastq > one_file_R1.gz
$ cat XXX_XXXXXX_L001_R2_001.fastq XXX_XXXXXX_L002_R2_001.fastq > one_file_R2.gz

[Brian Bushnell]

Merge them in the same order to keep odd things from happening later on. If you trim, use programs that will keep the mates in sync.

$ cat XXX_XXXXXX_L001_R1_001.fastq XXX_XXXXXX_L002_R1_001.fastq > one_file_R1[B].fq[/B]
$ cat XXX_XXXXXX_L001_R2_001.fastq XXX_XXXXXX_L002_R2_001.fastq > one_file_R2[B].fq[/B]

(I adjusted the file name extentions)

[sbdk82]

Thanks again !!

[ronton]

In addition to what others have stated, there is an arguement for not merging the files first, namely parallel computing. Each 'core' can process each smaller file in whichever way you wish, to be merged later in the pipeline. But, you have to keep track of what you are doing and it can be complicated.

Another note, I have had some strange results merging .fastq files in the same way you describe. For whatever reason, when I used Linux cat, and when I used Windows Powershell to merge the files, the resulting file just kept growing indefinitely and/or did not match the size of file 1 + file 2. However, when I used Windows copy /b <source1> + <source2> [....] <targetfile> the output file was as expected. I don't know why.

[Brian Bushnell]

If you do this:

cat *1.fastq > merged1.fastq

it will grow forever, because "merged1.fastq" matches *1.fastq and will also be used as an input. It should not happen if you explicitly name all of the input files.

[ronton]

Perhaps I used the wildcard, I will double check, thank you.