Olivar multiplex PCR tiling design

Description

Olivar is a Python software for multiplex PCR tiling design. Olivar first builds an index for each target of interest, incorporating undesired sequence features such as homologous regions, SNPs and extreme GC content. Olivar then designs tiled amplicons covering a single index or multiple indexes, and minimizes primer dimers with the SADDLE algorithm. Olivar is published as an article on Nature Communications.

Web Interface

A web interface is available at olivar.rice.edu, although it does not support all available functions at the moment.

Install with Bioconda (Linux x64 or Mac Intel chip)

1. Install Miniconda if not installed already (quick command line install)

2. Create a new Conda environment named "olivar" and install Olivar via Bioconda

conda create -n olivar olivar --channel conda-forge --channel bioconda --channel defaults --strict-channel-priority

[!TIP] Setting channel priority is important for Bioconda packages to function properly. You may also persist channel priority settings for all package installation by modifying your ~/.condarc file. For more information, check the Bioconda documentation.

3. Activate the new Conda environment and run Olivar

conda activate olivar
olivar --help

Dependencies

python >=3.8
blast >=2.12.0
biopython
numpy <2
pandas
plotly >=5.13.0
tqdm

Reproducibility

To reproduce the results in example_output (primers used in the publication), specify package versions during installation and run example.py.

conda create -n olivar olivar=1.1.5 blast=2.13.0 numpy=1 --channel conda-forge --channel bioconda --strict-channel-priority

[!CAUTION] Git LFS is needed to clone the example BLAST database. Without Git LFS, blastn won't run on the incomplete example BLAST database and Olivar will raise IndexError: list index out of range.

Usage

Input files

• (Required) Reference sequence for each target for tiling, in FASTA format (example). Ambiguous bases are not supported and may raise errors.

• (Optional) A list of sequence variations to be avoided for each reference, in csv format (example). Column "START" and "STOP" are required, "FREQ" is considered as 1.0 if empty. Other columns are not required. Coordinates are 1-based.

• (Optional) A BLAST database of non-specific sequences. More details can be found in Prepare a BLAST database.

[!NOTE] Coordinates are always 1-based, closed intervals, except fot the output .primer.bed file, which is in BED format.

Command-line interface

Olivar CLI comprises of four sub-commands: build, tiling, save and validate. Descriptions of command-line arguments can be found in Command-line parameters.

[!TIP] build, tiling, and validate support multiprocessing with -p option.

1. Build Olivar reference

A reference sequence in FASTA format is required, coordinates of sequence variations and BLAST database are optional. Only the first FASTA record is considered.

olivar build example_input/EPI_ISL_402124.fasta -v example_input/delta_omicron_loc.csv -d example_input/Human/GRCh38_primary -o example_output -p 1

An Olivar reference file (EPI_ISL_402124.olvr) will be generated, named by the ID of the FASTA record by default. Use multiple CPU cores (-p) to accelerate this process.

If you have multiple targets, run olivar build on each FASTA file and place all output .olvr files in the same directory.

In this step, the input reference sequence is chopped into kmers, and GC content, sequence complexity and BLAST hits are calculated for each kmer. Sequence variations are also labeled if coordinates are provided. A risk score is assigned to each nucleotide of the reference sequence, guiding the placement of primer design regions.

2. Design tiled amplicons

Input a single Olivar reference file generated in step 1, or a directory of multiple .olvr files (needs version ≥ 1.2). Set random seed (--seed) to make the results reproducible. Use multiple CPU cores (-p) to accelerate this process. Output files are listed below (coordinates are 1-based).

olivar tiling example_output/EPI_ISL_402124.olvr -o example_output --max-amp-len 420 --min-amp-len 252 --check-var --seed 10 -p 1

"olivar-design" is the name of the whole design (might contain multiple targets), and "EPI_ISL_402124" is the name of a single target, determined by the ID of the reference FASTA record by default (see step 1).

In this step, the placement of primer design regions (PDRs) is optimized based on the risk array (Fig.1d), and primer candidates are generated by SADDLE for each PDR in the optimized PDR set. SADDLE also minimizes primer dimer by exploring different combinations of primer candidates.

(Optional) Load from a previous Olivar design and save output files

Output files in step 2 can be generated repeatedly as long as the Olivar deisng file (.olvd) is provided.

olivar save example_output/olivar-design.olvd -o example_output

[!WARNING] .olvr and .olvd files are generated with pickle. Do NOT load those files from untrusted sources.

(Optional) Validate existing primer pools

Input should be a csv file, with four required columns: "amplicon_id" (amplicon name), "fP" (sequence of forward primer), "rP" (sequence of reverse primer) and "pool" (primer pool number, e.g., 1). This could be an Olivar designed primer pool generated in step 2, or primer pools that are not designed by Olivar. Output files are listed below (coordinates are 1-based). Use multiple CPU cores (-p) to accelerate this process.

olivar validate example_output/olivar-design.csv --pool 1 -d example_input/Human/GRCh38_primary -o example_output -p 1

Import Olivar as a Python package

Olivar can also be imported as a Python package, comprising of four functions with the same names and parameters as the four sub-commands in the CLI.

from olivar import build, tiling, save, validate

Refer to example.py for more details.

Command-line parameters

sub-command: build

olivar build fasta-file [--var <string>] [--db <string>] [--output <string>] 
[--title <string>] [--threads <int>]

sub-command: tiling

olivar tiling olvr-path [--output <string>] [--title <string>] [--max-amp-len <int>] 
[--min-amp-len <int>] [--w-egc <float>] [--w-lc <float>] [--w-ns <float>] [--w-var <float>] 
[--temperature <float>] [--salinity <float>] [--dg-max <float>] [--min-gc <float>] 
[--max-gc <float>] [--min-complexity <float>] [--max-len <int>] [--check-var] 
[--fp-prefix <DNA>] [--rp-prefix <DNA>] [--seed <int>] [--threads <int>]

sub-command: save

olivar save olvd-file [--output <string>]

sub-command: validate

olivar validate csv-file [--pool <int>] [--db <string>] [--output <string>] 
[--title <string>] [--max-amp-len <int>] [--temperature <float>] [--threads <int>]

Prepare a BLAST database

[!TIP] All BLAST related commands/scripts are installed along with Olivar.

• To make your own BLAST database with the makeblastdb command, check out the NCBI BLAST User Manual. \ The example BLAST database is created with 23 Chromosomes and MT of human genome assembly GRCh38, with the command (BLAST version 2.12.0):

  makeblastdb -in GRCh38_primary.fasta -dbtype nucl -title GRCh38_primary -parse_seqids -hash_index -out GRCh38_primary -max_file_sz 4GB -logfile makeblastdb.out -taxid 9606

• To download a pre-built BLAST database from NCBI (e.g., RefSeq representative gennomes for viruses), use the update_blastdb.pl script:

  update_blastdb.pl --decompress ref_viruses_rep_genomes

  For more details about update_blastdb.pl, check the BLAST Help. \ For more pre-built databases, check the NCBI FTP site.