Convert Bam To Fastq With Pe & Se Reads

Question

[Pasta]

Hi there,

EDIT

I filtered my PE reads based on quality. Filtering generated 3 kind of files:

• 1 file with PE of good quality
• 1 file with discarded reads,
• 1 file with orphan PE reads of good quality(unpaired PE).

I concatenated good quality reads, and submitted my reads to BWA to generate a SAM file. So basically, all my reads "look like" PE reads, even though I have unpaired reads.

!EDIT

I would like to convert a BAM/SAM file to FASTQ. The thing is that my BAM file contains Paired-end reads and Single-end reads, and I'd like a tool that will generate 3 files :

• 1 file with paired end reads #1
• 1 file with paired end reads #2
• 1 file with single end reads (unpaired)

I tried Picard and Bam2Fastq but this option doesnt exist. Any idea ?

Thanks

Answer 1

[Swbarnes2]

You could use samtools view to filter the .bam into three files; reads that are read 1, reads that are read 2, and reads from SE experiments. Picard can sort each .bam by read name, before you convert them to fastq.

[Swbarnes2]

What's so onerous about running samtools view 3 times to split up the .bam into those three groups?

samtools view -bf 64 mixed.bam > read1.bam samtools view -bf 128 mixed.bam > read2.bam samtools view -bF 1 mixed.bam > SE.bam

Then run bam2fastq on each of those files.

[Pasta]

Question edited with more details

Answer 2

[Shicheng Guo]

Swbarnes2 method is quite right. three Step:

1, Get all the reads belong to read1 (paired-end). With -f 64, you can get all the reads including 64, such as 65, 193 and so on

samtools view -bf 64 mixed.bam > read1.bam

2, Get all the reads belong to read2 (paired-end). With -f 128, you can get all the reads including 128, such as 129, 177 and so on

samtools view -bf 128 mixed.bam > read2.bam

3, Get all the reads belong to single-end. With -F 1, you can exclude all the pair-end reads.

samtools view -bF 1 mixed.bam > SE.bam

Answer 3

[Zev.Kronenberg]

Is this thread what you are looking for?

[Pasta]

Thanks. I read this thread already but I need to sort PE and SE while converting to fastq

[Zev.Kronenberg]

Do the read ids contain information on pairing? If so there are easy ways to do what you want. Can you provide a few read ids?