Genome Annotation

Lecture


After you have de novo assembled your genome sequencing reads into contigs, it is useful to know what genomic features are on those contigs. The process of identifying and labelling those features is called genome annotation.

Prokka is a "wrapper"; it collects together several pieces of software (from various authors), and so avoids "re-inventing the wheel".

Prokka finds and annotates features (both protein coding regions and RNA genes, i.e. tRNA, rRNA) present on on a sequence. Prokka uses a two-step process for the annotation of protein coding regions: first, protein coding regions on the genome are identified using Prodigal; second, the function of the encoded protein is predicted by similarity to proteins in one of many protein or protein domain databases. Prokka is a software tool that can be used to annotate bacterial, archaeal and viral genomes quickly, generating standard output files in GenBank, EMBL and gff formats. More information about Prokka can be found here.

Input data

Prokka requires assembled contigs. You can prepare you working directory for this annotation tutorial.

mkdir ~/annotation
cd ~/annotation

You will download an improved assembly of Mycoplasma genitalium into you data directory:

curl -O -J -L https://osf.io/7eaky/download

You will also need a proteins set specific of Mycoplasma for the annotation. Here is a file containing the Mycoplasma proteins retrieved from Swiss-Prot database (3041 sequences)

curl -O -J -L https://osf.io/xjm3n/download

Running prokka

prokka --outdir annotation --kingdom Bacteria \
--proteins uniprot_mycoplasma_reviewed.faa m_genetalium_improved.fasta

Once Prokka has finished, examine each of its output files.

Visualising the annotation

Artemis is a graphical Java program to browse annotated genomes. Download it here and install it on your local computer.

Copy the .gff file produced by prokka on your computer, and open it with artemis.

You will be overwhelmed and/or confused at first, and possibly permanently. Here are some tips: