4.1 Reference preparation¶

Purpose¶

Reference preparation builds all files needed for read classification, alignment, quantification, RPF density construction, and codon-level Ribo-seq analysis.

The workflow is divided into five steps:

Step 1: Prepare directories and reference files
Step 2: Build Bowtie indexes (genome, mRNA, rRNA, tRNA, ncRNA)
Step 3: Build the normalized RiboParser reference (rpf_Reference)
Step 4: Build the STAR index
Step 5: Build the RSEM index

All commands are run inside the 1.reference directory created in Step 1.

Inputs¶

Input	Description	Required
genome FASTA	genomic sequence	yes
GTF/GFF3 annotation	gene, transcript, exon, CDS, and UTR annotation	yes
feature table	optional NCBI feature table	no
external tools	Bowtie, STAR, RSEM, gffread	yes

Step 1: Prepare directories and reference files¶

1.1 Create directories¶

mkdir -p ./sce/1.reference/
cd ./sce/1.reference/

mkdir cdna genome gtf mrna ncrna rrna trna norm rsem-index star-index

1.2 Download reference files¶

Download the genome sequence, annotations, and optional feature table from NCBI. Use the URLs corresponding to your species.

# genome sequence
wget https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/146/045/GCF_000146045.2_R64/GCF_000146045.2_R64_genomic.fna.gz
# GTF or GFF3 annotation
wget https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/146/045/GCF_000146045.2_R64/GCF_000146045.2_R64_genomic.gtf.gz
wget https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/146/045/GCF_000146045.2_R64/GCF_000146045.2_R64_genomic.gff.gz
# optional feature table
wget https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/146/045/GCF_000146045.2_R64/GCF_000146045.2_R64_feature_table.txt.gz

gunzip *.gz

Step 2: Build Bowtie indexes¶

2.1 Generate cDNA from the genome and GFF3¶

The transcript FASTA used for mRNA/rRNA/tRNA/ncRNA extraction is generated from the genome and GFF3 with gffread, not downloaded directly.

gffread \
  -g ./GCF_000146045.2_R64_genomic.fna \
  ./GCF_000146045.2_R64_genomic.gff \
  -F \
  -w ./cdna.fa

2.2 Build the genome index¶

bowtie-build \
  ./GCF_000146045.2_R64_genomic.fna \
  ./genome/genome \
  --threads 12 \
  &>> ./genome/genome_build.log

2.3 Build the mRNA index¶

grep -i 'gbkey=mRNA' ./cdna.fa | cut -d ' ' -f 1 | cut -c 2- > ./mrna/mrna.ids

retrieve_seq \
  -i ./cdna.fa \
  -n ./mrna/mrna.ids \
  -o ./mrna/mrna.fa \
  &>> ./mrna/mrna_build.log

bowtie-build ./mrna/mrna.fa ./mrna/mrna --threads 12 &>> ./mrna/mrna_build.log

2.4 Build the rRNA index¶

grep -i 'gbkey=rRNA' ./cdna.fa | cut -d ' ' -f 1 | cut -c 2- > ./rrna/rrna.ids

retrieve_seq \
  -i ./cdna.fa \
  -n ./rrna/rrna.ids \
  -o ./rrna/rrna.fa \
  &>> ./rrna/rrna_build.log

bowtie-build ./rrna/rrna.fa ./rrna/rrna --threads 12 &>> ./rrna/rrna_build.log

2.5 Build the tRNA index¶

grep -i 'gbkey=tRNA' ./cdna.fa | cut -d ' ' -f 1 | cut -c 2- > ./trna/trna.ids

retrieve_seq \
  -i ./cdna.fa \
  -n ./trna/trna.ids \
  -o ./trna/trna.fa \
  &>> ./trna/trna_build.log

bowtie-build ./trna/trna.fa ./trna/trna --threads 12 &>> ./trna/trna_build.log

2.6 Build the ncRNA index¶

grep -iE 'gbkey=ncRNA|gbkey=lnc_RNA|gbkey=miRNA|gbkey=snoRNA|gbkey=snRNA|gbkey=misc_RNA' ./cdna.fa \
  | cut -d ' ' -f 1 \
  | cut -c 2- \
  > ./ncrna/ncrna.ids

retrieve_seq \
  -i ./cdna.fa \
  -n ./ncrna/ncrna.ids \
  -o ./ncrna/ncrna.fa \
  &>> ./ncrna/ncrna_build.log

bowtie-build ./ncrna/ncrna.fa ./ncrna/ncrna --threads 12 &>> ./ncrna/ncrna_build.log

Step 3: Build the normalized RiboParser reference¶

rpf_Reference normalizes the GTF/GFF3 annotation into RiboParser-compatible transcript tables and sequences, which are required by most downstream modules (e.g. quality control, gene-level analysis, and codon-level analysis).

3.1 Command¶

rpf_Reference \
  -g ./GCF_000146045.2_R64_genomic.fna \
  -t ./GCF_000146045.2_R64_genomic.gff \
  -u 30 \
  -o ./norm/gene \
  &>> ./norm/norm_build.log

3.2 Required parameters¶

Parameter	Meaning
`-g`, `--genome`	genome FASTA file
`-t`, `--gtf`	GTF or GFF3 annotation file
`-o`, `--output`	output prefix

3.3 Optional parameters¶

Parameter	Meaning
`-u`, `--utr`	add pseudo UTR length for leaderless transcripts
`-c`	retain only protein-coding transcripts
`-l`	retain only the longest protein-coding transcript per gene
`-w`	output full message table

3.4 Outputs¶

File	Description
`gene.norm.gtf`	normalized GTF
`gene.norm.txt`	RiboParser transcript annotation table
`gene.norm.rna.fa`	normalized transcript FASTA
`gene.norm.cds.fa`	normalized CDS FASTA

Step 4: Build the STAR index¶

The STAR index is built from the genome sequence and the normalized GTF, and is used for RNA-seq alignment.

STAR \
  --genomeSAindexNbases 11 \
  --runThreadN 12 \
  --runMode genomeGenerate \
  --genomeDir ./star-index \
  --genomeFastaFiles ./GCF_000146045.2_R64_genomic.fna \
  --sjdbGTFfile ./norm/gene.norm.gtf

Step 5: Build the RSEM index¶

The RSEM index is used for transcript-level quantification.

rsem-prepare-reference \
  -p 12 \
  --gtf ./norm/gene.norm.gtf \
  ./GCF_000146045.2_R64_genomic.fna \
  ./rsem-index/rsem

Quality checklist¶

Bowtie indexes exist for genome, mRNA, rRNA, tRNA, and ncRNA.
STAR index directory is complete.
RSEM index files are complete.
gene.norm.txt, gene.norm.rna.fa, and gene.norm.cds.fa are generated.
Chromosome names are consistent across genome, GTF/GFF3, BAM, and normalized reference.