roprokka 1.0.2

An add on function to consolidate Prokka output

Report On PROKKA

Prokka is a commonly used tool to annotate bacterial genomes. Roprokka (Row-pro-ka) is an add-on tool that gathers information from and computes metrics on Prokka output. Further, Roprokka will align common species identifier genes against the NCBI database and return the most similar species. The information is consolidated into one convenient results file to quickly assess the status of a bacterial genome assembly. Note that this tool does not run Prokka or comment on the significance of results.

About this tool

Prokka is a bioinformatics tool for annotating bacterial, archaeal, and vial genome assemblies. Output files are generated that contain a lot of information, including the annotation list, annotation sequences (nucleic acid and amino acid), and assembly statistics (Seemann, 2014). The information can be time consuming to gather when there are many samples, but the standard output format of the Prokka software makes it a good candidate for an automated mining system.

Getting Started

Requirements

• Prokka output. Before applying this tool, run Prokka.
• pysam
• blastn 2.10.1+
• python 3.7+

Installation

Roprokka is available on PyPI and can be installed using pip.

pip install roprokka

Usage

As input, this script takes a path to the directory of Prokka output.

Example:

roprokka -b path/to/blastn -i input_directory -o ouput_directory

TL;DR

Basic Assembly Stats

Displays the statistics presented in the Prokka .tsv file. Note, the statistics reported in this section are entirely dependendent upon the Prokka run parameters. For example, non-coding RNAs (ncRNAs) will only be reported if Prokka was run with the --rfam parameter.

Statistic	Description
contigs	Number of contigs or segments in the assembly
bases	Length of assembly, including all contigs
gene	Number of open reading frames identified
CDS	Number of coding sequences identified
rRNA	Number of ribosomal RNAs identified
tRNA	Number of transfer RNAs identified
misc_RNA	Number of miscellaneous RNAs identified
tmRNA	Number of transfer-messenger RNAs identified

Note that the CDS and RNA classes are subsets of the genes class.

Annotatios by function

The percent hypothetical is an important statistic for determining assembly quality. Due to limitations in current knowlege, there are many bacterial CDS that have not been annotated. Thus, an assembly where 40-60% of CDS are hypothetical proteins may still be a high quality assembly. However, if the percent hypothetical exceeds 90%, the quality/usability should be considered.

Statistic	Description
CDS	Number of coding sequences identified
hypothetical protein	Number of CDS classified as hypothetical
putative protein	Number of CDS classified as putative
perc_hypothetical	% of CDS classified as hypothetical
perc_putative	% of CDS classfied as putative

tRNA Breakdown

The number of tRNAs in an assembly can be indicative of assembly completeness and quality. A high quality assembly will have at least one (1 to many) tRNA for each of the 20 AAs (amino acids) (Land, 2014). A quick check of assembly completeness can be done (by looking at the 'tRNA AA range' line) to make sure that each AA is represented by at least one tRNA; however, note that it is not uncommon for an assembly to contain multiple tRNAs for some AAs. If all AAs have 2 or more corresponding tRNAs, consider checking your sample for multiple genomes. See supplementary.md for additional information.

A table of tRNAs by codon is also provided in the report file. It is common the number of tNRAs to be unevenly distributed across the codons. This could be due to reasons that are biological (codon usage bias) or computational (limited resolution for tRNA detection).

Example Codon Table

Ala:3	Arg:4	Asn:2	Asp: 1	Cys: 2
Gln:2	Glu:2	Gly:3	His:1	Ile:1
Leu:5	Lys:2	Met:3	Phe:1	Pro:3
Ser:4	Thr:3	Trp:1	Tyr:1	Val:3

Most importantly, Roprokka displays the range of AA that are represented tRNA AA range: 1-5

Number of Identifier Genes

The 16S rRNA and rpoB genes are commonly used for bacterial species identification because these genes are ubiquitous in bacteria and are somewhat species specific. An isolate may contain multiple copies of the 16S gene but bacteria typically only contain one copy of the rpoB gene. Multiple copies of the rpoB gene could indicate sequence duplications or multiple genomes.

BLAST Alignments

Manually finding and extracting these sequences from the .ffn file then running a BLAST alignment for each sequence can be labor and time intensive. Thus, Roprokka extracts the sequence of each identifier gene and aligns the sequence to the BLASTn database using blastn. Returned are the top 5 BLAST hits that have at least 90% sequence identity and at least 95% coverage of the query sequence. Each hit will have the following field information:

Field	Description
qseqid	query sequence ID
stitle	subject title
pident	% identity
qcovs	% query coverage per subject
qcovhsp	% query coverage per hsp
length	length of alignment
evalue	e value

The entirety of the BLAST results are returned to the user. This is because a single query might have multiple close matches to related species in the BLAST database. Sometimes distinguishing between these relults are better left to human the human eye.

The sequences for each identifier gene are retained in the output directory in case additional alignments are needed.

References (APA)

• Land, M. L., Hyatt, D., Jun, S. R., Kora, G. H., Hauser, L. J., Lukjancenko, O., & Ussery, D. W. (2014). Quality scores for 32,000 genomes. Standards in genomic sciences, 9(1), 1-10.
• Seemann, T. (2014). Prokka: rapid prokaryotic genome annotation. Bioinformatics, 30(14), 2068-2069.

Acknowledgements

• Cody Glickman