We produced a test ATAC-Seq data with 26.8 million 87bp paired-end reads using Chinese bulbul's feathers. I used Bowtie2 for the alignment and ATACseqQC for the quality analysis. The below are (1) fragment size distribution and (2) library complexity produced by ATACseqQC. I think that we have to optimize the ratio of cell number and Tn5 enzyme concentration. However, I don't know whether I have to increase or decrease Tn5 enzyme concentration. Also, does the library complexity result means that we have to increase the sequencing depth? Any suggestions are very welcome.