Thank you, @isaacovercast and @dereneaton ! I think that easy to operate should be ensured before any additional new function. So there is a kind of balance between how easy to do and how many functoins. And I would choice the operability as primary criteria and then is how many functions the software has. I don't remember how many lanes of data, while I know a single file, like SAMPLER1.fq.gz, is 742 MB. I tested a dataset on a HPC, the dataset containing five species and each species with R1 and R2 files. Whole assembly of this dataset costed about 8 hours, using 50 cores to run. My final dataset will contain more 100 species, so I try to know how can I decrease the time spent. Many thanks to both you two for your great creation and kindly helps.