mpa_vJan25_CHOCOPhlAnSGB_202503.fna not downloaded in the .tar file

Microbial community profiling MetaPhlAn
1

Hello,

I’ve been recommended your tool and I am looking forward to using it. However, I am having issues downloading the 2025 database.

My command was metaphlan --install --index mpa_vJan25_CHOCOPhlAnSGB_202503 --db_dir metaphlan.

This downloads the files

mpa_vJan25_CHOCOPhlAnSGB_202503.1.bt2l mpa_vJan25_CHOCOPhlAnSGB_202503.pkl
mpa_vJan25_CHOCOPhlAnSGB_202503.2.bt2l mpa_vJan25_CHOCOPhlAnSGB_202503.rev.1.bt2l
mpa_vJan25_CHOCOPhlAnSGB_202503.3.bt2l mpa_vJan25_CHOCOPhlAnSGB_202503.rev.2.bt2l
mpa_vJan25_CHOCOPhlAnSGB_202503.4.bt2l mpa_vJan25_CHOCOPhlAnSGB_202503_VINFO.csv
mpa_vJan25_CHOCOPhlAnSGB_202503.nwk mpa_vJan25_CHOCOPhlAnSGB_202503_VSG.fna.

I have also tried downloading directly from Index of /biobakery4/metaphlan_databases and the same file is missing.

My error is Thu Apr 2 17:06:21 2026: [Error] No Metaphlan database found at: /home/melody/dbs/metaphlan/mpa_vJan25_CHOCOPhlAnSGB_202503.fna to build the minimap index. I assume the file mpa_vJan25_CHOCOPhlAnSGB_202503.fna should be included in the download?

Looking forward to your response.

Best wishes

Melody

2

HI @melparker101
Yes, sorry about this there’s a bug that cause the removal of the *.fna file at installation, while it should be kept to be able to build the minimap index when you want to profile long-read sequencing data. We will post an update to remove this issue at some point. For now as a workaround you can concatenate the *_SGB.fna.bz2 and *_VSG.fna.bz2 file, that you find in the tar you have already located, into a file called mpa_vJan25_CHOCOPhlAnSGB_202503.fna (which is what metaphlan install does). Let me know if this works and make sure that when you run metaphlan you’re pointing to the directory where the database is located.

3

I followed these instructions and it seems to be working - just a note that the .mmi file is 56GB so I think it would be worth flagging somewhere in the database documentation that users planning on this approach are going to need more space than the 30GB mentioned for the database.

1 Like