Skip to content

Seed-based Gap Removal

Remove gaps relative to a reference (seed) sequence.

Usage

gapclean -i alignment.fa -o output.fa -s INDEX

Where INDEX is the 0-based position of your seed sequence.

How It Works

GapClean looks at your chosen seed sequence and removes any column where the seed has a gap. This effectively "projects" the alignment onto your reference sequence.

Example:

Original alignment:
>seed (index 0)
ATCG--ATCG
>seq2
ATCGATATCG
>seq3
ATCG--GTCG

Result with -s 0:
>seed
ATCGATCG
>seq2
ATCGATCG
>seq3
ATCGGTCG

Columns 4 and 5 (where seed has --) are removed.

Choosing a Seed Sequence

Use the First Sequence

gapclean -i alignment.fa -o output.fa -s 0

When to use: When your first sequence is a high-quality reference.

Use a Reference Genome

If you know sequence 3 is your reference: 

gapclean -i alignment.fa -o output.fa -s 2

When to use: When you have a known reference sequence.

Use the Longest Sequence

First, identify which sequence has the fewest gaps, then use it as seed.

When to use: When you want maximum coverage.

Common Use Cases

Comparing Against a Reference

# Your reference is the first sequence
gapclean -i query_vs_ref.fa -o cleaned.fa -s 0

This is perfect for: - Variant calling - Mutation analysis - Comparing sequences to a reference genome

Extracting Conserved Positions

# Use a well-characterized sequence as seed
gapclean -i alignment.fa -o conserved.fa -s 0

Site-wise Mutational Analysis

# Reference is first, variants follow
gapclean -i variants.fa -o positions.fa -s 0

Now each column corresponds to a position in your reference sequence.

Example Workflow

Step 1: Identify Your Reference

# Look at your FASTA file
head alignment.fa

>ReferenceStrain_NC_12345
ATCGATCG...
>StrainA
ATCGAT-G...
>StrainB
ATC-ATCG...

The reference is at index 0 (first sequence).

Step 2: Clean Alignment

gapclean -i alignment.fa -o cleaned.fa -s 0

Step 3: Verify

# Check that the first sequence has no gaps
grep -A 1 "^>ReferenceStrain" cleaned.fa

Tips

  • Seed index is 0-based (first sequence = 0, second = 1, etc.)
  • Choose a seed with few gaps for maximum coverage
  • The seed sequence will have no gaps in the output
  • This mode is deterministic: same seed = same result
  • Perfect for analyses where position matters

Comparison with Threshold Mode

Threshold Mode Seed Mode
Removes based on gap percentage Removes where seed has gaps
Position-independent Position matters (relative to seed)
Works on all sequences Focuses on one reference
Better for general cleaning Better for variant analysis