Total number of trimmed reads less than number of mapped reads

Question

[kspata]

Hi All,

I have forward and reverse trimmed reads. The total number of forward and reverse trimmed reads calculated using fastqc was 9,834,492 (with R1 and R2 being 4,917,246).

I aligned these reads to a multi fasta file using BWA v0.7.17. The following mapping statistics were obtained

9912451 + 0 in total (QC-passed reads + QC-failed reads)
0 + 0 duplicates
9863607 + 0 mapped (99.51%:-nan%)
9912451 + 0 paired in sequencing
4964874 + 0 read1
4947577 + 0 read2
9656111 + 0 properly paired (97.41%:-nan%)
9862604 + 0 with itself and mate mapped
1003 + 0 singletons (0.01%:-nan%)
4187 + 0 with mate mapped to a different chr
2507 + 0 with mate mapped to a different chr (mapQ>=5)

The number of total reads (9,912,451) is greater the sum of forward and reverse trimmed reads (9,834,492)

I checked if there were any secondary alignments (not primary alignments) using following command but found 0 secondary alignments

samtools view -f 256 -c Sample.bam

I checked number of unmapped reads with command samtools view -f 0x4 -c Sample.bam and found 48844 unmapped reads, this seems correct as

total (9912451) -mapped (9863607) = unmapped (48844)

I checked for supplementary alignments and found 77959 supplementary alignments

samtools view -f 2048 -c Sample.bam

if I subtract supplementary alignments count from mapped (9863607-77959 = 9834492) This number corresponds to the total number of trimmed reads (R1+R2)

So for read 1, 4,964,874 - 4,917,246 = 47,628, 47,628 forward reads were supplementary alignments

And for read 2, 4,947,577 - 4,917,246 = 30,331 , 30,331 reverse reads were supplementary alignments

I am confused as to how the read1 is greater than number of forward reads and same for reverse reads as well. Does this mean that a read which is primary alignment can also be considered as supplementary alignment?

Please let me know where I am going wrong.

Thanks in advance !!