How can I restore paired-end fastq files from a sorted bam file with fastq remain sorted?

Question

[lghust2011]

I have two paired-end FASQ files named fq1.fastq and fq2.fastq. Then I use BWA MEM to map these two files with the paired-end mode, like this:

bwa mem reference.fasta fq1.fastq fq2.fastq > result.sam

Then I transfer this sam file to bam format and sort it by coordinates with samtools:

samtools view result.sam > result.bam && samtools sort result.bam -o result_sort.bam

Now I want to restore these two fastq files from result_sort.bam. I know that I can use picard SamToFastq to do this. But reads in fastq files not remain sorted like the result_sort.bam. So, is there any other way that I can restore fastq files with reads sorted? I want to do this, because when I get fq1.fastq and fq2.fastq from result_sort.bam and use BWA to map then I again, I can get a sam file within that many reads are sorted. Any reply will be much appreciated!

[finswimmer]

I want to do this, because when I get fq1.fastq and fq2.fastq from result_sort.bam and use BWA to map then I again, I can get a sam file within that many reads are sorted

samtools sort accept also a sam file as input and can write to sam file. But why do you need a sam file instead of bam?

fin swimmer

[lghust2011]

Either sam or bam format is OK for me

[Gabriel R.]

May I ask why you need this?

[lghust2011]

I want to get the sorted fastq files, so I can get the sorted sam file directly when I run BWA MEM with sorted fastq files. But it seems very difficult!

[Gabriel R.]

Ok, what about reads getting assigned to random positions due to poor mapping quality, they will not end up in the same position? And to re-iterate my initial question, may I ask why you need this?

Answer 1

[Carlo Yague]

You can't do that !

In bwa manual, it is stated that :

bwa mem [...] db.prefix reads.fq [mates.fq]
[...] If mates.fq file is absent and option -p is not set, this command regards input reads are single-end. If mates.fq is present, this command assumes the i-th read in reads.fq and the i-th read in mates.fq constitute a read pair.

Meaning that your read pairs must be sorted by name in the fastq files inputed in bwa mem, not by coordinates.

---

EDIT : To clarify, the OP's idea of keeping the order of reads extracted from a coordinate-sorted bam file could work, but only with single-end reads. I guess that is goal is to avoid sorting the bam file after re-mapping (why would he want to re-map is another question) and saving some time. However, as stated above, aligners such as bwa-mem require the read pairs to correspond in the fastq.1 and fastq.2 files, making the coordinate-sorting inappropriate.

[lghust2011]

Thanks for your answer!