plannotate-plasmid-annotation

pLannotate Plasmid Annotation

Overview

pLannotate annotates plasmid sequences by running BLAST searches against a curated library of over 5,000 features sourced from Addgene, NCBI, and fpbase. It identifies promoters, terminators, antibiotic resistance genes, origins of replication, tags, and fluorescent proteins while correctly handling circular plasmid topology — avoiding split-feature artifacts that arise from naive linear alignment. Results are written as annotated GenBank files for downstream use in SnapGene, Benchling, or BioPython, as interactive HTML plasmid maps for sharing and review, and as CSV tables for programmatic filtering. Both a Python API and a command-line interface are provided; a Streamlit web app is also bundled for exploratory use.

When to Use

• Annotating a plasmid sequence received from a collaborator or downloaded from Addgene with no accompanying map
• Verifying that all expected elements (promoter, insert, resistance marker, origin) are present after assembly or mutagenesis
• Preparing a GenBank submission or Addgene deposit that requires a complete feature table
• Batch-annotating a library of synthetic constructs produced by combinatorial cloning
• Generating a shareable interactive plasmid map (HTML) without requiring SnapGene or Benchling licenses
• Checking a de-novo synthesized gene block for unintended regulatory elements or cryptic ORFs before cloning
• Use SnapGene or Benchling instead when you need a full-featured GUI plasmid editor with primer design and cloning simulation workflows; pLannotate is best for automated, scriptable annotation
• Use Prokka instead when annotating a complete bacterial genome or a large linear chromosomal sequence; pLannotate is optimized for plasmid-sized sequences up to ~50 kb

Prerequisites

• Python packages: plannotate, biopython (optional, for GenBank parsing)
• System dependency: BLAST+ must be available on PATH (installed automatically via conda; manual install needed for pip)
• Input: Plasmid sequence in FASTA format or as a plain Python string
• Data requirements: Sequences typically 1–20 kb; very large plasmids (>50 kb) may be slow

# Install via pip (requires BLAST+ on PATH)
pip install plannotate

# Install via conda (recommended — handles BLAST+ automatically)
conda install -c conda-forge -c bioconda plannotate

# Verify installation
plannotate --help
python -c "import plannotate; print('plannotate OK')"

Quick Start

from plannotate import annotate, write_genbank, create_bokeh_chart
from Bio import SeqIO

# Load plasmid from FASTA
record = next(SeqIO.parse("plasmid.fasta", "fasta"))
sequence = str(record.seq)

# Annotate (circular, against Addgene database)
results = annotate(sequence, linear=False, db="addgene")
print(f"Found {len(results)} features")
print(results[["Feature", "Feature_type", "pct_identity", "pct_query_cov"]].to_string())

# Export GenBank file
write_genbank(sequence, results, output_file="plasmid_annotated.gb")

# Generate interactive HTML map
create_bokeh_chart(sequence, results, output_file="plasmid_map.html")
print("Outputs: plasmid_annotated.gb, plasmid_map.html")

Workflow

Step 1: Load Plasmid Sequence

Load the plasmid sequence from a FASTA file, a GenBank file (stripping existing annotations for re-annotation), or a raw sequence string. Validate length and base composition before annotation.

from Bio import SeqIO
import os

# Option A: Load from FASTA
def load_fasta(path):
    record = next(SeqIO.parse(path, "fasta"))
    seq = str(record.seq).upper()
    return seq, record.id

# Option B: Load from GenBank (strip annotations, keep sequence)
def load_genbank(path):
    record = next(SeqIO.parse(path, "genbank"))
    seq = str(record.seq).upper()
    return seq, record.id

# Option C: Raw sequence string
raw_seq = "ATGCGTAAAGGAGAAGAACTTTTCACTGGAGTTGTCCCAATTCTTGTTGAATTAGATGGTGATGTT"

# Validate sequence
def validate_plasmid(seq, name="plasmid"):
    valid_bases = set("ATGCNRYSWKMBDHV")
    invalid = set(seq.upper()) - valid_bases
    if invalid:
        raise ValueError(f"Invalid bases in {name}: {invalid}")
    if len(seq) < 100:
        raise ValueError(f"Sequence too short ({len(seq)} bp); minimum 100 bp")
    gc = (seq.count("G") + seq.count("C")) / len(seq) * 100
    print(f"{name}: {len(seq):,} bp, GC={gc:.1f}%")
    return seq

seq, plasmid_id = load_fasta("plasmid.fasta")
validate_plasmid(seq, plasmid_id)

Step 2: Run BLAST-Based Annotation

Run annotation using the selected database. The linear flag controls whether the sequence is treated as circular (default for plasmids) or linear (for gene blocks and linear fragments).

from plannotate import annotate

# Annotate circular plasmid against the Addgene database (most comprehensive for common vectors)
results = annotate(
    seq,
    linear=False,       # False = circular plasmid (default)
    db="addgene",       # Database: "addgene", "fpbase", or "snapgene"
)

print(f"Total features detected: {len(results)}")
print(f"\nColumns: {list(results.columns)}")

# Preview feature table
cols = ["Feature", "Feature_type", "start", "end", "strand", "pct_identity", "pct_query_cov"]
print(results[cols].sort_values("start").to_string(index=False))

Step 3: Filter Features by Quality Thresholds

Review annotation confidence using BLAST identity and query coverage scores. High-confidence annotations have >95% identity and >90% coverage; partial hits may indicate truncated or mutated features.

import pandas as pd

# Inspect hit quality distribution
print("Identity percentile summary:")
print(results["pct_identity"].describe().round(1))
print("\nCoverage percentile summary:")
print(results["pct_query_cov"].describe().round(1))

# Separate high- and low-confidence hits
high_conf = results[
    (results["pct_identity"] >= 95) &
    (results["pct_query_cov"] >= 90)
].copy()

low_conf = results[
    (results["pct_identity"] < 95) |
    (results["pct_query_cov"] < 90)
].copy()

print(f"\nHigh-confidence features (identity>=95%, coverage>=90%): {len(high_conf)}")
print(f"Low-confidence / partial features:                       {len(low_conf)}")

if not low_conf.empty:
    print("\nLow-confidence features (review manually):")
    print(low_conf[["Feature", "Feature_type", "pct_identity", "pct_query_cov"]].to_string(index=False))

# Save filtered table
results.to_csv("all_features.csv", index=False)
high_conf.to_csv("high_confidence_features.csv", index=False)
print("\nSaved: all_features.csv, high_confidence_features.csv")

Step 4: Export Annotated GenBank File

Write the annotated sequence to GenBank format for import into plasmid editors (SnapGene, Benchling, Geneious, ApE) and for BioPython-based downstream analysis.

from plannotate import write_genbank

# Write full annotation (all features)
write_genbank(seq, results, output_file="plasmid_annotated.gb")
print("Written: plasmid_annotated.gb")

# Write with high-confidence features only
write_genbank(seq, high_conf, output_file="plasmid_highconf.gb")
print("Written: plasmid_highconf.gb")

# Verify using BioPython
from Bio import SeqIO
record = next(SeqIO.parse("plasmid_annotated.gb", "genbank"))
print(f"\nGenBank verification:")
print(f"  Sequence length: {len(record.seq):,} bp")
print(f"  Features: {len(record.features)}")
for feat in record.features:
    label = feat.qualifiers.get("label", ["(unlabeled)"])[0]
    print(f"  [{feat.type:20s}] {label} @ {feat.location}")

Step 5: Generate Interactive HTML Visualization

Create a Bokeh-based interactive plasmid map. The HTML file is self-contained and can be shared without any server infrastructure.

from plannotate import create_bokeh_chart

# Generate interactive circular plasmid map
create_bokeh_chart(
    seq,
    results,
    output_file="plasmid_map.html",
)
print("Interactive map saved: plasmid_map.html")
print("Open in any browser