When analyzing RNA-Seq data, rRNA and tRNA reads can be removed from the sequencing files. Here, I briefly describe how to do this step using ERNE

Prepare software

Download software from https://sourceforge.net/projects/erne/files/2.1.1/

Unzip and move erne-create and erne-filter to ~/local/bin

Install Seqkit conda install -c bioconda seqkit to use rmdup that remove duplicated fasta files.

Prepare rRNA and tRNA sequence

rRNA

rRNA sequences were downloaded from Silver

wget https://www.arb-silva.de/fileadmin/silva_databases/current/Exports/SILVA_128_LSUParc_tax_silva.fasta.gz

tRNA

tRNA sequences were downloaded from GtRNAdb.

wget http://gtrnadb2009.ucsc.edu/download/tRNAs/GtRNAdb-all-tRNAs.fa.gz

Combine two fasta files together and remove duplicate

Combine two fasta files as contaminate_rna.fa.

cat GtRNAdb-all-tRNAs.fa SILVA_128_LSUParc_tax_silva.fasta > contaminate_rna.fa
cat contaminate_rna.fa |seqkit rmdup -o contaminate_rna_uniq.fa

Prepare allign file

erne-create --output-prefix contaminate_rna --fasta contaminate_rna_uniq.fa & # takes about three hours

Run erne-filter to remove rRNA/tRNA

erne-filter --contamination-reference contaminate_rna.ebh --threads 20 --query1 test_trimmed.fq --output-prefix rmcontac&

As a result, the clean file rmcontac_1.fastq has 55167445 sequences, while the original file has 55301662 sequences. Over 99.76% of reads retained. This step is probably more important for small RNA-sequencing.