I want to use MEGAN for my microbiome study. I have illumina reads, which I have assembled into contigs. I want to run MEGAN on my reads/contigs.
I am sort of unsure what should I use. I can run MEGAN6 on my small illumina reads or MEGAN-LR on my assembled contigs. Which is the better option for me?
I have read that the MEGAN-LR is specifically written for long reads produced from PacBio and other technologies. However, my assembled contigs are in the range of 300nt-286,000 nt. How sensitive MEGAN-LR would be on this long size range?
All depends on the goals of your project. The short reads provide you with more interpretable measurements of abundance since assembly loses that information. On the other hand, for the long reads, MEGAN will be able to assign taxonomy and gene/function prediction better. However, that depends on the quality of your assembly. So I would suggest to try both and compare the results. If you want to save the computational time, you can troubleshoot on the subset of samples.