Let me explain the approach to handle your chloroplast contamination issue in your rice root metagenomic data. The key is to use the right reference sequences for filtering. You should start with the complete Oryza sativa chloroplast genome (NCBI Reference: NC_001320.1), specifically focusing on the 16S rRNA gene region within this sequence. However, using just the rice chloroplast sequence alone might not be comprehensive enough. A more thorough approach would be to incorporate the SILVA database's chloroplast-specific sequences, particularly from their SSU Ref NR 99 database, filtering for sequences with taxonomy strings containing "chloroplast." Alternatively, the PhyloFlash database offers well-curated chloroplast sequences that are particularly good at distinguishing between true Cyanobacteria and organelles. For the most effective results, I recommend creating a custom database that combines the rice chloroplast 16S sequence with known plant chloroplast sequences from SILVA, while keeping true Cyanobacterial sequences separate. This approach allows you to first remove plant-specific chloroplast reads before analyzing the remaining sequences for true Cyanobacteria, giving you a cleaner dataset for your analysis.

Thank you for your answer Sakshi Balyan , that's very helpful. I use SILVA database too, but I found a lot of sekuens that group either Cyanobacteria or chloroplast or Pseumonadota. I tried to use all of the as ecluding sequence, but I lost most of my data (I only found 14 bacteria species!). Which sequence should I use?