Off topic:Larger number of reads are aligned than input reads

Question

[SSK]

Hi!

I filtered my row reads by Trimmomatic, and aligned those by bwa mem. After that, I count input reads number with 'echo $(gzcat filtered_paird_fastq.gz | wc -l)/4|bc', and sum of the answers was ' 2736865184' .

Also, I count aligned reads number by using ' samtools flagstat bam'. The answer was like below.

2737435510 + 0 in total (QC-passed reads + QC-failed reads) 0 + 0 secondary 4886994 + 0 supplementary 0 + 0 duplicates 2737435510 + 0 mapped (100.00% : N/A) 2730564232 + 0 paired in sequencing 1365282116 + 0 read1 1365282116 + 0 read2 2270271490 + 0 properly paired (83.14% : N/A) 2730564232 + 0 with itself and mate mapped 0 + 0 singletons (0.00% : N/A) 194256760 + 0 with mate mapped to a different chr 47051705 + 0 with mate mapped to a different chr (mapQ>=5)

I found from this result that QC-passed reads ( line 1 ) is larger than input reads.

So, I want to ask what is the cause of this?

Thanks.