PrimerDesigner

End-to-end CRISPR primer design pipeline for non-model organisms.

Just provide a GenBank accession and a vector .dna file — BSgenome building, Bowtie indexing, gRNA library generation, off-target analysis, Golden Gate / Gibson cloning primer design, deletion arm primers, and SnapGene-compatible GenBank output are all handled in a single function call.

Features

Any organism — Works with any NCBI genome accession. BSgenome + Bowtie index are auto-built on first run (no separate setup)
Single function — design_grna_and_deletion() handles everything: genome download → BSgenome → Bowtie → gRNA → primers → Excel → GenBank
Genome-wide gRNA library — Composite scoring (GC%, Tm, position, off-target), methylation site filtering with IUPAC support
Off-target analysis — CrisprVerse crisprBowtie::runCrisprBowtie integration, verified for custom BSgenome
Vector-based oligo annealing — Reads actual vector sequence, scans ±20bp for Type IIS enzyme cut sites on both strands, computes real overhangs + fill sequences (not hardcoded)
9 Type IIS enzymes — BbsI, BpiI, BsaI, Eco31I, BsmBI, Esp3I, SapI, BspQI, PaqCI with NEB-verified cut patterns
Cut pattern polarity — Handles both 5' overhang (a < b, e.g. BbsI 2/6) and 3' overhang (a > b) with automatic oligo structure adjustment
Gibson Assembly — Tm-optimized overlap primers for inverse PCR amplification
Deletion arm primers — 4-primer Gibson design for upstream/downstream homology arms, circular genome support
Combined construct — All-in-one vector: gRNA spacer + deletion arms in a single construct
GenBank output — Annotated .gbk files for SnapGene (spacer, arms, primer binding sites with Tm)
Excel output — Multi-sheet .xlsx with unified primer tables and conditional formatting
SnapGene .dna support — Direct reading of SnapGene binary vector files via reticulate
Resume mode — Interrupt and restart long genome-wide runs without losing progress
Shared primer design across genomes — shared_primer_design() picks one UF/UR/DF/DR oligo set that works across multiple closely related target genomes, falling back to subgroup-splitting per strain only when no shared candidate exists

Installation

# 1. Bioconductor dependencies
if (!requireNamespace("BiocManager", quietly = TRUE))
  install.packages("BiocManager")

BiocManager::install(c(
  "crisprVerse", "crisprBase", "crisprDesign", "crisprBowtie",
  "BSgenome", "Biostrings", "GenomicFeatures", "GenomeInfoDb",
  "GenomicRanges", "Rsamtools", "rtracklayer", "Rbowtie"
))

# 2. Install from GitHub
devtools::install_github("JAEYOONSUNG/PrimerDesigner")

Python dependency (for SnapGene .dna files):

pip install snapgene_reader

Quick Start

Just provide a GenBank accession (or a local .gbff file) and a vector .dna file. Genome download, BSgenome package build, and Bowtie index creation are all handled automatically.

library(PrimerDesigner)

result <- design_grna_and_deletion(
  genbank_accession    = "GCF_030376765.1",        # NCBI accession or local .gbff/.gb/.gbk file path
  locus_tags           = c("QT235_RS00005", "QT235_RS00010"),  # single, multiple, or "all"
  nuclease             = "GeoCas9",
  grna_vector_file     = "~/vectors/pG1Kt-GeoCas9EF.dna",
  grna_start           = 8811,
  grna_end             = 8840,
  grna_cloning_method  = "golden_gate",
  grna_enzyme          = "BbsI",
  methylation_patterns = c("GCCAT", "CCANNNNNTTG"),
  output_file          = "all_primers.xlsx",
  output_dir           = "constructs"
)

# Access results
result$grna_primers      # gRNA cloning primers (2 per gRNA)
result$deletion_primers  # deletion arm primers (4 per locus_tag)
result$unified_primers   # all primers in unified long-format table

genbank_accession — accepts an NCBI accession (e.g. "GCF_030376765.1", auto-downloads) or a local GenBank file path (e.g. "~/genomes/my_genome.gbff"). Local .gbff / .gb / .gbk files are auto-detected; FASTA is extracted internally and multi-contig genomes are fully supported.

locus_tags — "QT235_RS00005" for a single gene, c("QT235_RS00005", "QT235_RS00010") for multiple genes, or "all" to run genome-wide gRNA + deletion primer design for every CDS.

What happens automatically on first run:

Download .fna / .gbff from NCBI (if not present)
Build + load BSgenome package
Build Bowtie index
Generate gRNA library + off-target analysis
Select best gRNA(s) + design cloning primers
Design deletion arm primers
Export Excel + GenBank output

On subsequent runs, pre-built BSgenome/Bowtie are reused automatically.

Multi-Genome Shared Primer Design

For knocking out the same gene / ortholog across multiple closely related strains, shared_primer_design() picks primers that bind every target genome simultaneously and falls back to per-strain subgroups only where cross-genome sharing is biologically impossible. All outputs land in a single 3-sheet Excel workbook plus an optional single-page summary PDF.

# 1) Discover the target CDS across every genome in a folder.
targets <- find_target_across_genomes(
  genbank_dirs = "path/to/gbk",
  query        = "TIGR02679",        # locus_tag, gene name, or product keyword
  query_type   = "product"
)

# 2) End-to-end design: shared UF/UR/DF/DR + gRNA cloning + check primers +
#    GenBank constructs + Excel workbook.
res <- shared_primer_design(
  target_table   = targets[, c("genome_id", "genbank_file", "locus_tag")],
  nuclease       = "GeoCas9",
  overlap_policy = "strict",         # inner primers pinned at 0 bp
  upstream_bp    = 500, downstream_bp = 500,
  min_arm_bp     = 150L, max_arm_bp   = 3000L,
  grna_vector_file  = "pJET.gb",
  grna_start        = 7823, grna_end  = 7850,
  combined_vector_file          = "pJET.gb",
  combined_grna_start           = 7823, combined_grna_end      = 7850,
  combined_deletion_start       = 3977, combined_deletion_end  = 4986,
  combined_construct_output_dir = "out/constructs",
  output_file                   = "out/design.xlsx"
)

# 3) One-page visual summary (arm alignment heatmaps, shared-primer bands,
#    flanking gene context, construct overview).
visualize_shared_design(
  result            = res,
  genbank_dir       = "path/to/gbk",
  construct_gbk_dir = "out/constructs",
  output_file       = "out/summary.pdf"
)

Key behaviours:

Strict overlap policy — Under overlap_policy = "strict" the inner primers (UR, DF) are hard-pinned to the deletion boundary (0 bp tolerance). If no shared candidate exists at the boundary, the selector subgroup-splits into per-strain clusters rather than silently drifting off the stop codon.
Shared core + subgroup split — Outer primers (UF, DR) and check primers progressively share across as many genomes as possible; strain-specific clusters are emitted only when biologically forced (e.g., transposon-adjacent targets).
Cross-role RC collision guard — The selector rejects any primer whose reverse complement equals its partner on the same arm (prevents the Gibson pair from collapsing the effective homology arm).
Colony-PCR check primers — cF / cR pairs are placed just outside the effective homology arm (check_outer_pad + check_search_window, default 50 bp + 800 bp) with a progressive window-widening loop: the window doubles up to check_search_window_max (default 6000 bp) until a single shared pair covers every genome, falling back to per-strain cR / cF clusters only when the surrounding context genuinely diverges. Tm is enforced at 55 ± 3 °C with a pair ΔTm ≤ 2 °C, strict genome-uniqueness, and a 6 bp 3'-complementarity heterodimer guard.
Per-genome WT / deletion PCR band sizes — For every (genome, cF, cR) triple the expected wild-type amplicon and post-deletion amplicon length are reported alongside delta_bp (equal to the deletion span), so colony PCR gel interpretation is one table lookup.
3-sheet Excel output:
1. primer_order — single combined list of every oligo to order (arm primers with Gibson overhangs, gRNA cloning primers with Type IIS overhangs, cF / cR check primers). A group column separates arm / gRNA / check, rows are shaded by used_by cluster, and thin borders automatically outline each role block.
2. check_primers — per-genome diagnostic PCR band sizes (WT, deletion, Δ) plus pair ΔTm and dimer flags.
3. final_construct_groups — canonical construct list with gRNA protospacer and the full UF/UR/DF/DR primer cluster assignment for each build.
Single-page summary PDF — visualize_shared_design() renders a 7-panel overview: construct maps, insert-site zoom, sharing matrix, oligo order list, upstream / downstream arm alignment heatmaps with cluster-banded primer lanes, and the flanking gene context (which CDSes occupy the homology arms on each genome).

The legacy lower-level design_shared_grna_and_deletion() is kept exported for back-compat; shared_primer_design() is the recommended entry point.

Gibson deletion library (batch knockout construct generation)

shared_primer_design() doubles as a Gibson deletion library builder: feed it the same target gene across N strains and it returns N combined deletion constructs in one pass, sharing every UF / UR / DF / DR primer that is biologically sharable and only diverging into strain-specific sub-clusters where the flanking sequence forces it. For every construct you get:

A native SnapGene .dna file written by the bundled in-package binary editor. Every feature color, custom DNA coloring stripe, and Primers-panel entry from the source plasmid round-trips unchanged. The designed UF / UR / DF / DR and gRNA-cloning primers are appended to the Primers panel under their canonical order names (<locus_tag>_UF_clstr<N>_F, sgRNA_<locus_tag>_clstr<N>_<inv|OA>_<F|R>), so they appear natively alongside the vector's existing primers rather than as generic GenBank features.
A parallel .gbk used for the PDF summary and downstream tooling that speaks GenBank.

Enable this with write_dna_copies = TRUE (it runs in addition to the GenBank writer):

res <- shared_primer_design(
  target_table                  = targets[, c("genome_id","genbank_file","locus_tag")],
  nuclease                      = "GeoCas9",
  overlap_policy                = "strict",
  upstream_bp                   = 500,  downstream_bp = 500,
  min_arm_bp                    = 150L, max_arm_bp    = 3000L,
  grna_vector_file              = plasmid,  grna_start = 7823, grna_end = 7850,
  grna_cloning_method           = "gibson",   # Gibson sgRNA clone (inv); use "golden_gate" for OA
  combined_vector_file          = plasmid,
  combined_grna_start           = 7823, combined_grna_end      = 7850,
  combined_deletion_start       = 3977, combined_deletion_end  = 4986,
  combined_construct_output_dir = "out/constructs",  # emits <label>_combined_construct.{gbk,dna}
  output_file                   = "out/design.xlsx",
  write_dna_copies              = TRUE
)

Sequence-based homolog fallback (for strains where annotation misses the target)

find_target_across_genomes(..., sequence_fallback = TRUE) (default on) first scans annotation fields, then for every genome that returned nothing it runs an AA-level MMseqs2 translated search against the genome nucleotide sequence, maps the hit back to the closest annotated CDS, and appends a row tagged match_field = "sequence_homolog". Reference strains whose ortholog is annotated only as "hypothetical protein" are therefore caught automatically. If MMseqs2 is not installed the fallback transparently switches to a Biostrings CDS-level local alignment (slower but dependency-free).

MMseqs2 install (recommended):

# Homebrew (macOS)
brew install mmseqs2
# or Bioconda (cross-platform)
conda install -c bioconda mmseqs2

The package auto-detects the binary across the common homebrew / miniforge / miniconda / anaconda install paths on load, so no manual PATH manipulation is needed. Status is reported on library(PrimerDesigner) startup; find_mmseqs_binary() returns the cached location at any time.

Run it from the terminal

Single command that uses the test genomes shipped with the package (inst/extdata/*.dna) and writes a complete Gibson deletion library + Excel + PDF to ~/Desktop/jetD_out/:

Rscript -e 'library(PrimerDesigner)
extdata <- system.file("extdata", package = "PrimerDesigner")
plasmid <- file.path(extdata, "plasmid/pG1Kt-GeoCas9EF-OA-HDVrbz-sfGFP-ACrec.dna")
out_dir <- "~/Desktop/PD_result"; construct_dir <- file.path(out_dir, "constructs")
dir.create(out_dir, recursive = TRUE, showWarnings = FALSE)

targets <- find_target_across_genomes(
  genbank_dirs = extdata, query = "TIGR02679", query_type = "product",
  interactive = FALSE, one_per_genome = FALSE, sequence_fallback = TRUE)

res <- shared_primer_design(
  target_table     = targets[, c("genome_id","genbank_file","locus_tag")],
  nuclease         = "GeoCas9", overlap_policy = "strict",
  upstream_bp      = 500, downstream_bp = 500,
  min_arm_bp       = 150L, max_arm_bp    = 3000L,
  primer_min_length = 20L, primer_max_length = 35L,
  grna_vector_file  = plasmid, grna_start = 7823, grna_end = 7850,
  grna_cloning_method = "gibson",
  combined_vector_file          = plasmid,
  combined_grna_start           = 7823, combined_grna_end      = 7850,
  combined_deletion_start       = 3977, combined_deletion_end  = 4986,
  combined_construct_output_dir = construct_dir,
  output_file                   = file.path(out_dir, "design.xlsx"),
  write_dna_copies              = TRUE)

visualize_shared_design(
  result           = res, genbank_dir = extdata,
  construct_gbk_dir = construct_dir,
  output_file      = file.path(out_dir, "summary.pdf"))'

The same block runs under source() from an R session or from Rscript my_design.R if you save it as a file. Replace "TIGR02679" / "gibson" / coordinate values with your own target / cloning method / vector coordinates.

Pipeline Overview

                    design_grna_and_deletion()
                              │
         ┌────────────────────┼────────────────────-┐
         │      All steps run automatically         │
         │                    │                     │
         ▼                    ▼                     ▼
  ┌───────────────┐ ┌─────────────────┐ ┌─────────────────┐
  │ Genome Setup  │ │ gRNA Library    │ │ Primer Design   │
  │ (auto-build)  │ │                 │ │                 │
  ├───────────────┤ ├─────────────────┤ ├─────────────────┤
  │ .fna download │ │ findSpacers     │ │ gRNA cloning    │
  │ .gbff download│ │ Off-target n0/n1│ │  ├─ Golden Gate │
  │ BSgenome build│ │ Composite score │ │  └─ Gibson      │
  │ Bowtie index  │ │ Methylation     │ │ Deletion arms   │
  └───────────────┘ │ filtering       │ │ (4-primer)      │
                    └─────────────────┘ │ Combined GenBank│
                                        │ Excel output    │
                                        └─────────────────┘

Internal call chain:

design_grna_and_deletion()
  └─ design_grna_construct()
       └─ run_gRNA_list_generator()
            ├─ download_genbank_fna()       # auto-download if missing
            ├─ download_genbank_gbff()      # auto-download if missing
            ├─ run_bowtie_build()           # auto-build if missing
            ├─ build_bsgenome_from_accession()  # auto-build if missing
            └─ findSpacers + off-target + scoring
  └─ design_deletion_primers()              # 4-primer Gibson deletion arms
  └─ write_combined_construct_genbank()     # SnapGene-compatible GenBank output

Golden Gate Oligo Annealing — Vector-Based Overhang Detection

Instead of using hardcoded overhangs, PrimerDesigner scans the actual vector sequence to locate Type IIS enzyme recognition sites and designs oligo annealing primers with real overhangs and fill sequences.

How it works

Load vector file (.dna / .gb) → scan ±20bp flanking regions around the stuffer
Search both strands → use only sites that cut toward the stuffer (warns on wrong-orientation sites)
Automatically determine overhang type from cut pattern (a/b):
- a < b (e.g. BbsI 2/6) → 5' overhang, length = b - a
- a > b (hypothetical 6/2) → 3' overhang, length = a - b, oligo structure auto-adjusted
- a == b → blunt end → incompatible with Golden Gate (warning)
Automatically include fill sequences between enzyme cut site and stuffer boundary

Supported Type IIS Enzymes

Enzyme	Recognition	Cut Pattern	Overhang	Source
BbsI	`GAAGAC`	(2/6)	4nt 5'	NEB R0539
BpiI	`GAAGAC`	(2/6)	4nt 5'	Thermo ER1011
BsaI	`GGTCTC`	(1/5)	4nt 5'	NEB R0535
Eco31I	`GGTCTC`	(1/5)	4nt 5'	Thermo ER0291
BsmBI	`CGTCTC`	(1/5)	4nt 5'	NEB R0739
Esp3I	`CGTCTC`	(1/5)	4nt 5'	NEB R0734
SapI	`GCTCTTC`	(1/4)	3nt 5'	NEB R0569
BspQI	`GCTCTTC`	(1/4)	3nt 5'	NEB R0712
PaqCI	`CACCTGC`	(4/8)	4nt 5'	NEB R0745

Oligo design formula

F oligo = [LEFT_OH] + [fill_L] + [G] + spacer + [fill_R]
R oligo = [RIGHT_OH] + RC(fill_R) + RC(spacer) + [C] + RC(fill_L)

Where LEFT_OH, RIGHT_OH, fill_L, fill_R are all derived from the actual vector sequence around the enzyme cut sites. For 3' overhang enzymes, the RC logic is automatically adjusted.

Usage modes

# Mode 1: Vector-based (recommended) — real overhangs from vector sequence
design_cloning_primers(gRNA_df, vector_file = "vector.dna",
                       start = 8811, end = 8840,
                       cloning_method = "golden_gate", enzyme = "BbsI")

# Mode 2: Default overhangs (no vector needed, e.g. BbsI = CACC/AAAC)
design_cloning_primers(gRNA_df, cloning_method = "golden_gate", enzyme = "BbsI")

# Mode 3: Custom overhangs
design_cloning_primers(gRNA_df, cloning_method = "golden_gate",
                       custom_overhangs = list(F_5prime = "CACC", R_5prime = "AAAC"))

Supported Nucleases

Nuclease	PAM	PAM Side	Spacer Length
GeoCas9	NNNNCAAA	3'	21 bp
SpCas9	NGG	3'	20 bp
FnCas12a	TTTV	5'	23 bp

Custom nucleases can be added via crisprBase::CrisprNuclease().

Key Functions

Main Pipeline

Function	Description
`design_grna_and_deletion()`	Full pipeline (single genome): accession → BSgenome → gRNA → cloning primers → deletion arms → Excel → GenBank. All steps automated
`shared_primer_design()`	Full pipeline (multi-genome): shared UF/UR/DF/DR across strains + gRNA cloning + colony-PCR check primers + combined construct GenBank + 3-sheet Excel. See Multi-Genome Shared Primer Design
`find_target_across_genomes()`	Resolve the same ortholog (by locus_tag / gene / product keyword) across a folder of genome files, with interactive ambiguity handling
`visualize_shared_design()`	Single-page 7-panel PDF summary for a `shared_primer_design()` result

Step-by-Step (call individually if needed)

Function	Description
`design_grna_construct()`	gRNA selection + cloning primers only (no deletion)
`run_gRNA_list_generator()`	gRNA library only (includes auto BSgenome/Bowtie build)
`design_cloning_primers()`	Golden Gate (OA) / Gibson (IA) primers only
`design_deletion_primers()`	4-primer deletion arm design (single gene)
`batch_deletion_primers()`	Multi-gene deletion primer batch

gRNA Library Utilities

Function	Description
`generate_gRNA_for_sequence()`	Generate gRNAs from a DNA sequence string
`generate_gRNA_for_locus_tags()`	Generate gRNAs for specific locus tags
`build_gRNA_library()`	Build gRNA library (scoring + GenBank merge)
`calculate_composite_score()`	Weighted scoring: GC%, Tm, position, off-target
`generate_gRNA_names()`	Standardized gRNA naming (e.g. `Geo_dnaA_g1`)
`filter_methylation_sites()`	IUPAC methylation motif filtering
`nuclease_with_parameter()`	Return nuclease object + PAM/spacer parameters

Genome Setup (called automatically — no manual call needed)

Function	Description
`build_bsgenome_from_accession()`	Build BSgenome (auto-called internally, bypasses UCSC validation)
`run_bowtie_build()`	Build Bowtie index (auto-called internally)
`download_genbank_fna()`	Download .fna (auto-called internally)
`download_genbank_gbff()`	Download .gbff (auto-called internally)
`forge_BSgenome()`	Low-level BSgenome package builder

File I/O

Function	Description
`read_vector_file()`	Read `.dna` / `.gb` / `.fasta` vector files
`read_genome_genbank()`	Parse genome `.gbff` → sequence + feature table
`export_constructs_to_dna()`	Batch `.gbk → .dna` via SnapGene CLI (single kill-once flow). Alternative to `write_dna_copies = TRUE`
`write_deletion_genbank()`	Write a SnapGene-compatible `.gbk` for a deletion construct
`write_grna_vector_genbank()`	Write a gRNA vector `.gbk` with annotated protospacer

MMseqs2 / Environment Helpers

Function	Description
`find_mmseqs_binary(refresh = FALSE)`	Locate and cache the MMseqs2 binary (scans PATH + homebrew + miniforge / miniconda / anaconda). Called automatically on load
`install_mmseqs_via_conda(env_name = NULL)`	One-liner installer: uses the detected `mamba` / `micromamba` / `conda` binary to install `mmseqs2` from bioconda into the base env or a named env, then refreshes the cache
`patch_snapgene_reader(verbose = TRUE)`	Rewrites the installed `snapgene_reader` Python module to (1) fix its "too many values to unpack" crash on multi-valued qualifiers and (2) parse the Primers panel (block 5) that upstream ignores. Called automatically on load
`write_grna_vector_genbank()`	GenBank output: spacer inserted in vector
`write_deletion_genbank()`	GenBank output: deletion construct
`format_primer_name()`	Customizable primer naming pattern

Cloning Methods

Golden Gate (Oligo Annealing)

Anneals two oligos and ligates directly into the sticky ends created by Type IIS restriction enzyme digestion. When a vector file is provided, overhangs and fill sequences are automatically computed from the actual cut pattern.

result <- design_grna_and_deletion(
  genbank_accession   = "GCF_030376765.1",
  locus_tags          = "QT235_RS00005",
  grna_vector_file    = "~/vectors/pG1Kt-GeoCas9EF.dna",
  grna_start          = 8811,
  grna_end            = 8840,
  grna_cloning_method = "golden_gate",
  grna_enzyme         = "BbsI",
  output_file         = "primers.xlsx",
  output_dir          = "constructs/"
)

Gibson Assembly (Inverse PCR)

Inserts the spacer via inverse PCR amplification of the vector using Tm-optimized overlap primers.

result <- design_grna_and_deletion(
  genbank_accession   = "GCF_030376765.1",
  locus_tags          = "QT235_RS00005",
  grna_vector_file    = "~/vectors/pG1Kt-GeoCas9EF.dna",
  grna_start          = 8811,
  grna_end            = 8840,
  grna_cloning_method = "gibson",
  tm_target           = 60,
  output_file         = "primers.xlsx",
  output_dir          = "constructs/"
)

Deletion (4-Primer Gibson)

Clones upstream/downstream homology arms into the vector via Gibson Assembly. Supports circular genomes. Automatically included in design_grna_and_deletion().

Output Files

output_dir/
├── all_primers.xlsx                          # Unified primer table (multi-sheet)
│   ├── Sheet: all_primers                    #   Long-format: gRNA + deletion primers
│   ├── Sheet: gRNA_library                   #   Full gRNA library with scores
│   └── Sheet: deletion_detail                #   Wide-format deletion arms
├── QT235_RS00005_g1_combined_construct.gbk   # GenBank: gRNA + arms in one vector
├── QT235_RS00010_g1_combined_construct.gbk
└── .grna_primers_cache.rds                   # Resume cache (auto-generated)

Open .gbk files in SnapGene to visually inspect spacer, homology arms, and primer binding sites.

Requirements

Category	Packages
R	>= 4.0
Bioconductor	crisprVerse, crisprBase, crisprDesign, crisprBowtie, BSgenome, Biostrings, GenomicFeatures, GenomeInfoDb, GenomicRanges, Rsamtools, rtracklayer, Rbowtie
CRAN	openxlsx, TmCalculator, glue, dplyr, reticulate, data.table, stringr, xmltodict (via reticulate)
System	Bowtie (installed via `Rbowtie`)
System (optional)	MMseqs2 — enables fast sequence-based homolog search in `find_target_across_genomes(..., sequence_fallback = TRUE)`. Without it, the function transparently falls back to a Biostrings CDS-level aligner. Install via `brew install mmseqs2`, `conda install -c bioconda mmseqs2`, or `install_mmseqs_via_conda()` from within R
Python	`snapgene_reader` + `biopython<1.82` (for `.dna` file support). `biopython` must be pinned below 1.82 because newer versions drop APIs `snapgene_reader` 0.1.23 depends on

Troubleshooting

UCSC seqlengths warning on first run

When building a BSgenome for non-model organisms, you may see warnings like:

Error in evaluating the argument 'x' in selecting a method for function 'seqlengths':
  UCSC library operation failed
BSgenome getSeq validation: FAILED -- will rebuild

This is expected and harmless. Non-UCSC genomes (i.e., most non-model organisms) do not have UCSC-style chromosome naming, so the initial seqlengths validation fails. PrimerDesigner automatically detects this and rebuilds the BSgenome package, which then works correctly. You can safely ignore this warning.

"MMseqs2 not found" on load

[PrimerDesigner] MMseqs2 not found — sequence-based homolog fallback will use the slower Biostrings engine.

Not an error — find_target_across_genomes(sequence_fallback = TRUE) transparently falls back to Biostrings::pairwiseAlignment over the annotated CDS table. To enable the 10–100× faster MMseqs2 path, install it via one of:

install_mmseqs_via_conda()                # auto-detects conda/mamba, installs into base env
install_mmseqs_via_conda(env_name = "mmseqs")  # install into a dedicated env

brew install mmseqs2                      # macOS
conda install -c bioconda mmseqs2         # cross-platform

PrimerDesigner auto-detects the binary across PATH, /opt/homebrew, miniforge3, mambaforge, miniconda3, anaconda3, and r-miniconda — no manual Sys.setenv(PATH=...) needed. Status is printed on library(PrimerDesigner) load; find_mmseqs_binary() returns the cached location.

"Cannot read SnapGene .dna file" / "too many values to unpack (expected 2)"

snapgene_reader 0.1.23 has two bugs that bite real-world .dna files:

It crashes on feature qualifiers that carry more than two V items.
It silently drops the Primers panel (block id 5), so every primer tracked in SnapGene's Primers side panel is lost on import.

PrimerDesigner patches the installed module in place on package load via patch_snapgene_reader(). If you installed snapgene_reader into a different Python env, call it manually after switching envs. The fix is idempotent.

Also pin Biopython: snapgene_reader 0.1.23 uses APIs removed in Biopython 1.82+:

pip install 'biopython<1.82'

"bundled dna_editor.py not found in PrimerDesigner package"

Only happens when calling via devtools::load_all() from a tree that hasn't been rebuilt after the editor was added. Either reinstall (R CMD INSTALL) or restart the R session; the fallback path resolver (find.package + inst/python/) in the wrapper usually catches this automatically.

~ tilde in output paths produces no .dna file

Fixed in .shared_write_combined_dna() as of commit f9f7a02 — R-side path.expand() runs before the Python writer opens the file. If you see this on an older install, pull master and reinstall.

Citation

If you use PrimerDesigner in your research, please cite:

Sung, J.-Y. (2025). PrimerDesigner: End-to-end CRISPR primer design pipeline
for non-model organisms. R package version 1.0.0.
https://github.com/JAEYOONSUNG/PrimerDesigner

Name		Name	Last commit message	Last commit date
Latest commit History 40 Commits
R		R
inst		inst
man		man
.Rbuildignore		.Rbuildignore
.gitignore		.gitignore
DESCRIPTION		DESCRIPTION
NAMESPACE		NAMESPACE
PrimerDesigner.Rproj		PrimerDesigner.Rproj
PrimerDesigner_for_Deletion.py		PrimerDesigner_for_Deletion.py
PrimerDesigner_for_Gibson.py		PrimerDesigner_for_Gibson.py
README.md		README.md

Folders and files

Latest commit

History

Repository files navigation

PrimerDesigner

Features

Installation

Quick Start

Multi-Genome Shared Primer Design

Gibson deletion library (batch knockout construct generation)

Sequence-based homolog fallback (for strains where annotation misses the target)

Run it from the terminal

Pipeline Overview

Golden Gate Oligo Annealing — Vector-Based Overhang Detection

How it works

Supported Type IIS Enzymes

Oligo design formula

Usage modes

Supported Nucleases

Key Functions

Main Pipeline

Step-by-Step (call individually if needed)

gRNA Library Utilities

Genome Setup (called automatically — no manual call needed)

File I/O

MMseqs2 / Environment Helpers

Cloning Methods

Golden Gate (Oligo Annealing)

Gibson Assembly (Inverse PCR)

Deletion (4-Primer Gibson)

Output Files

Requirements

Troubleshooting

Citation

License

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases 1

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages