Hello everyone,

I have a question about small RNA-seq analysis using Bowtie2 and featureCounts.

I aligned my reads with Bowtie2 using the -k 100 option, which allows Bowtie2 to report up to 100 valid alignment locations per read. Then I ran featureCounts using the default settings.

featureCounts \  
     -a "$GTF" \                                
     -o "$COUNT_OUT" \           
     -T "$THREADS" \                   
     -t miRNA \                               
     -g gene_id \                            
     -s 1 \    

I am trying to understand what happens to the multimapped reads in this case. With default featureCounts settings, are all multimapped reads discarded completely, even if Bowtie2 marks one alignment as the primary alignment? Or does featureCounts still count the primary alignment and ignore the secondary alignments?

When featureCounts is run in default mode after Bowtie2 -k 100, does the final count matrix include only uniquely mapped reads, or does it count one primary alignment from each multimapped read?

I read the featureCounts user guide, but I am still a bit confused about how multimapped reads are handled, especially when the alignments come from Bowtie2 using -k 100 or with other value of -K.