The number of reads from the BMA file is more than the number of reads from its original FASTQ file

Question

[Gary]

Hi,

I have a FASTQ file (name as Origin_ELW24.fastq.gz) with 107,856,041 single-end 75bp reads. After trimming and alignment, we get a BAM file (name as ELW24.bam) via STAR. I use commands below to convert the BMA file to a new FASTQ file (name as New_ELW24.fastq.gz). The number of reads in the new FASTQ file is 122,444,250 that more than the number of reads in the original FASTQ file. I don't understand why the number of reads will increase after the trimming and alignment. Could you help me? Many thanks.

bedtools bamtofastq -i ELW24.bam -fq New_ELW24.fastq
gzip -c New_ELW24.fastq > New_ELW24.fastq.gz

Best,

Gary

Answer 1

[Fatima]

bedtools bamtofastq:

each alignment in the BAM file is converted to a FASTQ record in the -fq file.

It's probably because each read (that is not uniquely mapped) can be mapped to the reference more than once (multi-mapping reads).

https://github.com/alexdobin/STAR/blob/master/doc/STARmanual.pdf https://bedtools.readthedocs.io/en/latest/content/tools/bamtofastq.html

Answer 2

[Arup Ghosh]

The extra number of reads you are observing because of the secondary alignment. Can you share the samtools flagstat results?