While working on the miRNA project, I encountered weird error with Picard, sam_bam convert and even remove duplicate. I thought that it would be caused from jave security problem. In the end, it turned out that it was cause by “duplicate miRBase IDs”

grep hsa-let-7c  /ddn/gs1/home/li11/refDB/miRBase/hsa_combined.fa
>hsa-let-7c MIMAT0000064 Homo sapiens let-7c
>hsa-let-7c MI0000064 Homo sapiens let-7c stem-loop

I have to replace all “spaces”.

Now, with WES project, I am using Picard to remove the duplicate. Here are some good postes

1. People are getting high duplication in WES
2. How to determine removed reads for PE sample
3. A thorough Picard documentation
4. Remove or not remove??
5. Duplication metric in details

Why Picard Markduplicate does not produce histogram

1. First post concludes that it is caused by multiple-groups

, but I do NOT think so.