GOAT

GOAT is a Linux-oriented PWM-based motif search pipeline for DNA FASTA files. It converts motif files into normalized PWMs, estimates fixed score thresholds, and scans sequences using a core-first strategy.

Requirements

• Linux
• Bash
• C++17 compiler, such as g++

On Ubuntu/Debian:

sudo apt update
sudo apt install -y build-essential

Build

bash compile.sh

This creates:

• meme2pwm
• get_thr
• search_motif

Quick Start

bash compile.sh
bash run.sh examples/motifs.txt examples/sequences.fa results/motif_hits.tsv

Outputs:

• results/thresholds.tsv
• results/motif_hits.tsv

Run

bash run.sh <motif_file> <sequence_file_or_dir> <results_file_or_dir> [options]

Required arguments:

• motif_file: MEME file or simple Motif: file.
• sequence_file_or_dir: FASTA file, or directory of .fa, .fasta, or .fna files.
• results_file_or_dir: output TSV file; output directory when sequence input is a directory.

Options:

Option	Default	Meaning
--threshold-file PATH	next to results	Threshold output TSV.
--core-length N	5	Length of the selected core window.
--full-pct PCT	95.0	Full motif threshold percentile.
--core-pct PCT	75.0	Core region threshold percentile.
--iterations N	10000	Simulation count per motif.
--precision N	6	Threshold decimal precision.
`--bayes 0	1`	0
--pseudocount N	45	Pseudocount value passed to meme2pwm.

--bayes 1 enables Bayesian PWM smoothing during motif normalization. This reduces zero-probability issues by blending each PWM row with a small prior. --pseudocount controls how strongly the original motif row is weighted relative to that prior.

Example with options:

bash run.sh motifs.txt seq.fa results/hits.tsv \
  --threshold-file results/thresholds.tsv \
  --core-length 6 \
  --full-pct 95 \
  --core-pct 75 \
  --iterations 20000

Batch Running

Batch mode is automatic when the sequence input is a directory:

bash run.sh motifs.txt fasta_dir results/batch_hits

The directory may contain files ending in .fa, .fasta, or .fna. GOAT writes one result file per FASTA input:

results/batch_hits/<input_stem>_motif_hits.tsv

The threshold file is written once and reused for all FASTA files:

results/batch_hits/thresholds.tsv

Batch mode also accepts the same threshold and PWM options:

bash run.sh motifs.txt fasta_dir results/batch_hits \
  --threshold-file results/batch_hits/thresholds.tsv \
  --core-length 6 \
  --bayes 1

Motif Input

Simple motif format:

Motif:example_motif
0.90 0.03 0.04 0.03
0.03 0.90 0.04 0.03
0.03 0.04 0.90 0.03
0.03 0.04 0.03 0.90

Each row is ordered as A C G T. Rows are normalized automatically. MEME files are also supported when they contain letter-probability matrix:.

Search Method

For each motif, get_thr selects the lowest-entropy contiguous core window. During search, search_motif first scores the core region at each candidate position. Only windows that pass the core threshold are scored against the full motif. Sequence scanning is parallelized across available CPU threads.

Because scores are defined as negative log-likelihoods, lower values indicate better motif matches.

Output

Threshold file:

# motif_id	full_threshold	core_threshold	core_start	motif_length

Motif hit file:

sequence_id	motif_id	strand	start	end	full_score	core_score	matched_sequence

Coordinates are 1-based and inclusive on the input FASTA sequence.

Manual Usage

./meme2pwm examples/motifs.txt /tmp/pwm.tsv 0 45
./get_thr /tmp/pwm.tsv results/thresholds.tsv 95.0 75.0 10000 5 6
./search_motif examples/sequences.fa /tmp/pwm.tsv results/thresholds.tsv results/motif_hits.tsv 5