🧠 Motif Discovery Fundamentals

Q: What is motif discovery?

Motif discovery is the computational identification of short, recurring, conserved sequence patterns (motifs) in DNA, RNA, or protein sequences. It's used to find transcription factor binding sites, regulatory elements, splice sites, and functional signatures that are over-represented in a set of sequences.

Q: What is a Position Weight Matrix (PWM)?

A Position Weight Matrix (PWM), also called a Position-Specific Scoring Matrix (PSSM), is a representation of a motif. It gives a score for each nucleotide (or amino acid) at each position of the motif, based on observed frequencies in a set of aligned binding sites, often with a background model correction.

Q: Which motif discovery tool should I use?

Use MEME for comprehensive de novo discovery, DREME or STREME for fast short motifs, HOMER for combined known + de novo discovery and annotation, and MEME-ChIP for ChIP-Seq peak analysis. For scanning known motifs, use FIMO (MEME suite) or the JASPAR database with tools like TFBSTools.

Q: What is the difference between de novo and known motif discovery?

De novo motif discovery finds motifs from scratch using only the input sequences (e.g., MEME, DREME, HOMER). Known motif discovery scans input sequences for matches to a database of known motifs (e.g., JASPAR, HOCOMOCO, CIS-BP), using tools like FIMO or MAST. They are complementary.

Q: What is a motif logo?

A motif logo (sequence logo) is a graphical representation of a motif where the height of each letter is proportional to its information content (conservation) at that position and the size of each letter reflects its frequency. It's the standard way to visualize PWMs and is produced by WebLogo, Skylign, or ggseqlogo.

🎯 What is Motif Discovery?

Motif discovery is the computational identification of short, recurring, conserved sequence patterns (motifs) in DNA, RNA, or protein sequences. It's used to find transcription factor binding sites, regulatory elements, splice sites, and functional signatures that are over-represented in a set of sequences. It's a cornerstone of regulatory genomics and epigenomics.

Why It Matters

TF Binding: Regulatory sites. Enhancers: Discovery. Splicing: Splice sites. Function: Protein motifs. Disease: Regulatory variants. Networks: Gene regulation.

📌 Sequence Patterns📌 Regulatory Elements📌 TF Binding📌 PWM

# Motif Discovery — the big picture
# Finding recurring patterns in sequences.

# CONCEPTUAL OVERVIEW
print("""
=== Motif Discovery ===

What it is:
  - Find recurring patterns
  - Short sequences (6-20 bp)
  - Over-represented
  - Functional sites

Why it matters:
  - TF binding sites
  - Enhancers
  - Promoters
  - Splice sites
  - Protein motifs
  - Regulatory networks

Types of motifs:
  - DNA motifs (TF binding)
  - RNA motifs (splicing, structure)
  - Protein motifs (domains, active sites)
  - Structural motifs

Representations:
  - Consensus sequence
  - PWM / PSSM
  - IUPAC codes
  - Regular expressions
  - HMM

Discovery methods:
  - Enumerative (DREME, STREME)
  - Probabilistic (MEME)
  - Bayesian (Gibbs)
  - Combinatorial (HOMER)

Workflow:
  1. Collect sequences
  2. Identify motifs (MEME, HOMER)
  3. Compare to databases (JASPAR)
  4. Scan genome (FIMO)
  5. Annotate
  6. Validate

Applications:
  - ChIP-Seq analysis
  - ATAC-Seq analysis
  - Enhancer discovery
  - Regulatory variants
  - Gene networks
  - Disease mechanisms
""")

# WHY IT MATTERS
print("=== Why It Matters ===")
print("""
Biological questions:
  - What TFs bind these regions?
  - What are the regulatory elements?
  - How is gene expression controlled?
  - What motifs are enriched in peaks?
  - How do variants affect binding?

Motif discovery answers:
  - Binding motifs
  - Regulatory logic
  - Networks
  - Functional sites
  - Disease mechanisms
""")

# HISTORICAL CONTEXT
print("=== Historical Context ===")
print("""
1980s: Consensus sequences
1990s: MEME, Gibbs sampling
2000s: HOMER, DREME, JASPAR
2010s: ChIP-Seq era, MEME-ChIP
2020s: Deep learning, single-cell
""")

# CHALLENGES
print("=== Challenges ===")
print("""
1. MOTIF LENGTH
   - Short motifs
   - Many false positives
   - Degenerate

2. BACKGROUND
   - Composition bias
   - Dinucleotide bias
   - Need good background

3. MULTIPLE MOTIFS
   - Co-occurring
   - Composite
   - Cooperative

4. REPEATS
   - Low complexity
   - Masking needed

5. VALIDATION
   - Functional testing
   - Conservation
   - ChIP-Seq
""")

🎯 Why Motif Discovery Matters

🎯 The Value of Motif Discovery

Motif discovery is the key to understanding gene regulation, interpreting ChIP-Seq peaks, and decoding regulatory variants.

Key Applications

ChIP-Seq: TF motifs. ATAC-Seq: Regulatory. Enhancers: Discovery. Variants: Impact. Networks: Regulation. Disease: Mechanisms.

🎯 ChIP-Seq🎯 Enhancers🎯 Variants🎯 Networks

# Why Motif Discovery Matters
print("=== 1. ChIP-Seq Analysis ===")
print("""
ChIP-Seq:
  - TF binding sites
  - Peaks
  - Motif discovery

Workflow:
  1. ChIP-Seq
  2. Peak calling
  3. Extract sequences
  4. Motif discovery (MEME, HOMER)
  5. Compare to JASPAR
  6. Validate

Examples:
  - p53 motifs
  - CTCF motifs
  - OCT4 motifs
  - NF-κB motifs

Tools:
  - MEME-ChIP
  - HOMER
  - ChIPseeker
""")

print("\n=== 2. Enhancer Discovery ===")
print("""
Enhancers:
  - Distal regulatory
  - H3K27ac peaks
  - TF motifs

Workflow:
  1. H3K27ac ChIP-Seq
  2. Peak calling
  3. Motif discovery
  4. TF assignment
  5. Networks

Examples:
  - Cell-type specific
  - Cancer enhancers
  - Super-enhancers

Tools:
  - HOMER
  - MEME
  - ROSE
""")

print("\n=== 3. Regulatory Variants ===")
print("""
Variants:
  - GWAS hits
  - eQTLs
  - Motif disruption

Workflow:
  1. Variant
  2. Reference motif
  3. Alternate motif
  4. Score change
  5. Impact

Tools:
  - FIMO
  - JASPAR
  - motifbreakR
  - atSNP
""")

print("\n=== 4. Gene Regulatory Networks ===")
print("""
Networks:
  - TF → target genes
  - Motif in promoters
  - Network inference

Workflow:
  1. ChIP-Seq / ATAC
  2. Motif discovery
  3. Target genes
  4. Network
  5. Analysis

Tools:
  - HOMER
  - MEME
  - RcisTarget
  - SCENIC
""")

print("\n=== 5. Splicing Regulation ===")
print("""
Splicing:
  - Splice sites
  - ESE/ESS
  - ISE/ISS

Motifs:
  - Donor: GT
  - Acceptor: AG
  - Branch point
  - Polypyrimidine tract

Tools:
  - MEME
  - SpliceAid
  - ESEfinder
  - MaxEntScan
""")

print("\n=== 6. Protein Motifs ===")
print("""
Protein motifs:
  - Domains
  - Active sites
  - Binding sites

Examples:
  - Walker A: GXXXXGK[ST]
  - Zinc finger: CXXC
  - N-glycosylation: NX[ST]
  - Nuclear localization: K[KR]X[KR]

Tools:
  - PROSITE
  - Pfam
  - ELM
  - MEME
""")

print("\n=== 7. RNA Motifs ===")
print("""
RNA motifs:
  - Binding sites
  - Structural
  - Regulatory

Examples:
  - RBP motifs
  - Riboswitches
  - miRNA seeds

Tools:
  - MEME
  - HOMER
  - RNAcontext
  - GraphProt
""")

print("\n=== 8. Disease Mechanisms ===")
print("""
Disease:
  - Regulatory variants
  - Enhancer hijacking
  - Motif disruption

Workflow:
  1. Variant
  2. Motif
  3. Impact
  4. Mechanism
  5. Treatment

Examples:
  - Cancer
  - Autoimmune
  - Neurodegenerative
""")

🧬 Types of Motifs

🎯 Many Kinds of Patterns

Motifs come in many forms: DNA, RNA, protein, and structural. Each has its own discovery methods.

Types

DNA: TF binding. RNA: Splicing, RBP. Protein: Domains. Structural: 3D. Regulatory: Enhancers. Consensus: Sequence.

🧬 DNA🧬 RNA🧬 Protein🧬 Structural

# Types of Motifs
print("=== 1. DNA Motifs ===")
print("""
DNA motifs:
  - TF binding sites
  - Regulatory elements
  - Promoters, enhancers
  - 6-20 bp

Examples:
  - TATA box: TATAAA
  - GC box: GGGCGG
  - CCAAT box: CCAAT
  - p53: RRRCWWGYYY
  - CTCF: CCCTC

Databases:
  - JASPAR
  - HOCOMOCO
  - CIS-BP
  - UniPROBE
""")

print("\n=== 2. RNA Motifs ===")
print("""
RNA motifs:
  - Splicing signals
  - RBP binding sites
  - Structural elements
  - miRNA targets

Examples:
  - 5' splice site: GURAGU
  - 3' splice site: YAG
  - Branch point: YNYURAY
  - miRNA seed: 2-8 nt

Tools:
  - MEME
  - HOMER
  - RNAcontext
  - GraphProt
""")

print("\n=== 3. Protein Motifs ===")
print("""
Protein motifs:
  - Domains
  - Active sites
  - Binding sites
  - Modification sites

Examples:
  - Walker A: GXXXXGK[ST]
  - Walker B: [RK]XXXXGXXXX[LIVMF]
  - Zinc finger: CXXC
  - N-glycosylation: NX[ST]
  - NLS: K[KR]X[KR]

Databases:
  - PROSITE
  - Pfam
  - ELM
  - InterPro
""")

print("\n=== 4. Structural Motifs ===")
print("""
Structural motifs:
  - 3D patterns
  - Folds
  - Supersecondary

Examples:
  - Helix-turn-helix
  - Zinc finger
  - Leucine zipper
  - β-hairpin
  - Greek key

Databases:
  - CATH
  - SCOP
  - PROSITE
""")

print("\n=== 5. Regulatory Motifs ===")
print("""
Regulatory motifs:
  - Enhancers
  - Promoters
  - Insulators
  - Silencers

Examples:
  - Enhancer: cell-type specific
  - Promoter: TATA, CpG
  - Insulator: CTCF
  - Silencer: REST

Tools:
  - HOMER
  - MEME
  - ChIPseeker
""")

print("\n=== 6. Composite Motifs ===")
print("""
Composite motifs:
  - Multiple TFs
  - Cooperative binding
  - Clusters

Examples:
  - NF-κB + IRF
  - STAT + GATA
  - OCT4 + SOX2
  - AP-1 + NFAT

Tools:
  - MCAST
  - Cluster-Buster
  - Cister
""")

print("\n=== 7. Repeat Motifs ===")
print("""
Repeat motifs:
  - Tandem repeats
  - STRs
  - Microsatellites

Examples:
  - (CAG)n: Huntington's
  - (CGG)n: Fragile X
  - (GAA)n: Friedreich's

Tools:
  - Tandem Repeats Finder
  - TRF
  - RepeatMasker
""")

print("\n=== 8. Consensus Sequences ===")
print("""
Consensus:
  - Most common base
  - IUPAC codes
  - Simple

IUPAC:
  A: A
  C: C
  G: G
  T: T
  R: A/G
  Y: C/T
  S: G/C
  W: A/T
  K: G/T
  M: A/C
  N: Any

Example:
  p53: RRRCWWGYYY
  Meaning:
    R = A/G
    W = A/T
    Y = C/T
""")

📊 PWM & PSSM

🎯 Representing Motifs Numerically

A Position Weight Matrix (PWM) is the standard numerical representation of a motif, giving a score for each base at each position.

PWM

PWM: Position weight. PSSM: Scoring matrix. Frequencies: Counts. Log-odds: Scores. Background: Correction. Logo: Visualization.

📊 PWM📊 PSSM📊 Log-odds📊 Information content

# PWM & PSSM
print("=== Position Weight Matrix ===")
print(""
PWM:
  - Position Weight Matrix
  - 4 x L (DNA)
  - 20 x L (protein)
  - Frequencies or log-odds

Example (DNA):
  Position:  1  2  3  4  5  6
  A:       0.1 0.6 0.1 0.1 0.1 0.1
  C:       0.1 0.1 0.1 0.6 0.1 0.1
  G:       0.7 0.2 0.1 0.2 0.1 0.1
  T:       0.1 0.1 0.7 0.1 0.7 0.7

Consensus: G A T C T T
""")

print("\n=== PSSM (Log-Odds) ===")
print(""
PSSM:
  - Position-Specific Scoring Matrix
  - Log-odds
  - Background correction

Formula:
  score(i,j) = log2( f(i,j) / b(j) )

Where:
  f(i,j) = frequency of base j at position i
  b(j) = background frequency

Example:
  f = 0.7, b = 0.25
  score = log2(0.7/0.25) = 1.49

  f = 0.1, b = 0.25
  score = log2(0.1/0.25) = -1.32
""")

print("\n=== Information Content ===")
print(""
Information content:
  IC = log2(4) - H
  H = Shannon entropy

Formula:
  H = -Σ p(i) log2 p(i)

Example:
  All A: H = 0, IC = 2 bits
  Random: H = 2, IC = 0 bits

Logo:
  - Height = IC
  - Letter size = frequency
  - Conserved = tall
  - Variable = short
""")

print("\n=== PWM Formats ===")
print(""
MEME:
  - Text format
  - Frequencies
  - With background

JASPAR:
  - PFM: Position Frequency Matrix
  - PWM: Log-odds
  - PFM → PWM

TRANSFAC:
  - Matrix format
  - Frequencies

HOMER:
  - Motif file
  - Frequencies
  - Log-odds
""")

print("\n=== Example MEME ===")
print(""
MEME version 5.0

ALPHABET= ACGT

strands: + -

Background letter frequencies
A 0.25 C 0.25 G 0.25 T 0.25

MOTIF p53
letter-probability matrix: alength= 4 w= 10 nsites= 100 E= 1e-50
  0.100  0.100  0.700  0.100
  0.600  0.100  0.200  0.100
  0.700  0.100  0.100  0.100
  0.100  0.600  0.100  0.200
  ...
""")

print("\n=== Example JASPAR ===")
print(""
>MA0001.1 AGL3
A  [ 0  3 79 40 66 48 65 11 65  0 ]
C  [94 75  4  3  1  2  5  2  3  3 ]
G  [ 1  0  3  4  1  0  5  3 28 88 ]
T  [ 2 19 11 50 29 47 22 81  1  6 ]
""")

print("\n=== Tools ===")
print(""
1. MEME
   - PWM generation
   - Discovery

2. JASPAR
   - Database
   - PFM/PWM

3. Biopython
   - Motif parsing
   - PSSM

4. TFBSTools (R)
   - PWM manipulation
   - Scanning

5. Logomaker (Python)
   - Logo visualization
""")

print("\n=== Best Practices ===")
print(""
1. Use log-odds
2. Correct background
3. Use pseudocounts
4. Check IC
5. Visualize logos
6. Validate
7. Report parameters
""")

⚙️ Discovery Algorithms

🎯 How Motifs Are Found

Motif discovery algorithms fall into several categories: enumerative, probabilistic, Bayesian, and combinatorial.

Algorithms

Enumerative: DREME. Probabilistic: MEME. Bayesian: Gibbs. Combinatorial: HOMER. Consensus: Simple. HMM: Profile.

⚙️ Enumerative⚙️ Probabilistic⚙️ Bayesian⚙️ Combinatorial

# Discovery Algorithms
print("=== 1. Enumerative ===")
print(""
Enumerative:
  - Enumerate all k-mers
  - Count occurrences
  - Statistical significance

Steps:
  1. Extract k-mers (k=6-10)
  2. Count in input
  3. Compare to background
  4. Score
  5. Report

Pros:
  - Fast
  - Complete
  - No local optima

Cons:
  - k limited
  - Degenerate motifs

Tools:
  - DREME
  - STREME
  - Weeder
  - YMF
""")

print("\n=== 2. Probabilistic ===")
print(""
Probabilistic:
  - PWM
  - Expectation-Maximization
  - Local optima

Steps:
  1. Initialize PWM
  2. E-step: assign sites
  3. M-step: update PWM
  4. Repeat until convergence

Pros:
  - Handles degeneracy
  - PWM output
  - Standard

Cons:
  - Local optima
  - Needs initialization
  - Multiple runs

Tools:
  - MEME
  - AlignACE
  - Consensus
""")

print("\n=== 3. Bayesian ===")
print(""
Bayesian:
  - Gibbs sampling
  - Posterior
  - Flexible

Steps:
  1. Initialize
  2. Sample motif positions
  3. Update PWM
  4. Repeat
  5. Posterior

Pros:
  - No local optima
  - Uncertainty
  - Flexible

Cons:
  - Slow
  - Convergence
  - Complex

Tools:
  - Gibbs Motif Sampler
  - BioProspector
  - MODELLER
""")

print("\n=== 4. Combinatorial ===")
print(""
Combinatorial:
  - Known motifs
  - De novo
  - HOMER

Steps:
  1. Enumerate
  2. Score
  3. Compare to database
  4. Report

Pros:
  - Fast
  - Known + de novo
  - Annotation

Cons:
  - k limited
  - Simple motifs

Tools:
  - HOMER
  - Pscan
  - Clover
""")

print("\n=== 5. Consensus ===")
print(""
Consensus:
  - Simple
  - Most common base
  - IUPAC

Methods:
  - Frequency
  - Threshold
  - IUPAC

Pros:
  - Simple
  - Fast
  - Interpretable

Cons:
  - Loses info
  - No scoring
  - Not quantitative

Tools:
  - Simple scripts
  - Consensus
""")

print("\n=== 6. HMM ===")
print(""
HMM (Hidden Markov Model):
  - Profile HMM
  - Probabilistic
  - Flexible

Steps:
  1. Build HMM
  2. Train
  3. Score
  4. Align

Pros:
  - Flexible
  - Indels
  - Probabilistic

Cons:
  - Complex
  - Slow

Tools:
  - HMMER
  - SAM
  - Profile HMM
""")

print("\n=== Algorithm Comparison ===")
print(f"{'Algorithm':<20} {'Speed':<15} {'Accuracy':<15} {'Tools':<20}")
print("-" * 70)
print(f"{'Enumerative':<20} {'Fast':<15} {'Good':<15} {'DREME':<20}")
print(f"{'Probabilistic':<20} {'Medium':<15} {'Good':<15} {'MEME':<20}")
print(f"{'Bayesian':<20} {'Slow':<15} {'Good':<15} {'Gibbs':<20}")
print(f"{'Combinatorial':<20} {'Fast':<15} {'Good':<15} {'HOMER':<20}")
print(f"{'Consensus':<20} {'Very fast':<15} {'Basic':<15} {'Scripts':<20}")
print(f"{'HMM':<20} {'Medium':<15} {'Excellent':<15} {'HMMER':<20}")

print("\n=== Choosing ===")
print(""
Choose algorithm:
  - Fast: DREME, STREME
  - Comprehensive: MEME
  - Known + de novo: HOMER
  - Long motifs: MEME
  - Short motifs: DREME
  - Protein: MEME, HMMER
  - Flexible: Gibbs
""")

🎯 MEME Suite

🎯 The Comprehensive Toolkit

The MEME Suite is the most comprehensive collection of motif discovery and analysis tools.

MEME

MEME: Discovery. DREME: Short motifs. STREME: Fast. FIMO: Scanning. MAST: Search. TOMTOM: Compare. MEME-ChIP: ChIP.

🎯 MEME🎯 DREME🎯 FIMO🎯 TOMTOM

# MEME Suite
print("=== MEME ===")
print(""
MEME:
  - De novo discovery
  - Probabilistic
  - PWM
  - Comprehensive

Command:
  meme input.fa \
    -dna \
    -oc meme_output \
    -nmotifs 10 \
    -minw 6 \
    -maxw 20 \
    -mod zoops

Options:
  -dna: DNA sequences
  -nmotifs: Number (10)
  -minw: Min width (6)
  -maxw: Max width (20)
  -mod: Distribution (zoops, anr, oops)

Output:
  meme.txt
  meme.html
  meme.xml
""")

print("\n=== DREME ===")
print(""
DREME:
  - Short motifs
  - Enumerative
  - Fast

Command:
  dreme -p input.fa -o dreme_output

Options:
  -p: Primary sequences
  -n: Background
  -e: E-value
  -m: Max motifs

Output:
  dreme.txt
  dreme.html
""")

print("\n=== STREME ===")
print(""
STREME:
  - Fast
  - Sensitive
  - Enumerative

Command:
  streme --p input.fa --o streme_output

Options:
  --p: Primary sequences
  --n: Background
  --thresh: Threshold

Output:
  streme.txt
  streme.html
""")

print("\n=== FIMO ===")
print(""
FIMO:
  - Scan sequences
  - PWM matching
  - P-values

Command:
  fimo motif.meme genome.fa > fimo.txt

Options:
  --thresh: P-value threshold
  --bgfile: Background
  --parse-genomic-coord

Output:
  fimo.txt
  fimo.html
""")

print("\n=== MAST ===")
print(""
MAST:
  - Motif search
  - Multiple motifs
  - E-values

Command:
  mast motif.meme genome.fa > mast.txt

Options:
  -mt: Motif threshold
  -ev: E-value

Output:
  mast.txt
  mast.html
""")

print("\n=== TOMTOM ===")
print(""
TOMTOM:
  - Compare motifs
  - Database matching
  - E-values

Command:
  tomtom query.meme database.meme -oc tomtom_output

Options:
  -thresh: Threshold
  -dist: Distance

Output:
  tomtom.tsv
  tomtom.html
""")

print("\n=== MEME-ChIP ===")
print(""
MEME-ChIP:
  - ChIP-Seq analysis
  - Peak sequences
  - Comprehensive

Command:
  meme-chip peaks.fa -oc meme_chip_output

Options:
  -db: Database (JASPAR)
  -meme-nmotifs: Number
  -dreme-e: E-value

Output:
  meme-chip.html
  combined.meme
""")

print("\n=== Other MEME Tools ===")
print(""
1. CentriMo
   - Central motif enrichment
   - ChIP-Seq

2. SpaMo
   - Spacing analysis
   - Composite motifs

3. GOMo
   - Gene ontology
   - Motif function

4. MCAST
   - Cluster analysis
   - Composite motifs

5. AME
   - Motif enrichment
   - Known motifs
""")

print("\n=== Workflow ===")
print(""
1. Extract sequences:
   bedtools getfasta -fi genome.fa -bed peaks.bed -fo peaks.fa

2. Discover motifs:
   meme peaks.fa -dna -oc meme_output

3. Compare to JASPAR:
   tomtom meme_output/meme.txt JASPAR2020.meme -oc tomtom

4. Scan genome:
   fimo meme_output/meme.txt genome.fa > fimo.txt

5. Enrichment:
   ame peaks.fa JASPAR2020.meme -o ame_output
""")

🏠 HOMER

🎯 Known + De Novo Discovery

HOMER (Hypergeometric Optimization of Motif EnRichment) is a popular tool for combined known and de novo motif discovery.

HOMER

Known: Database. De novo: Discovery. Fast: Efficient. Annotation: Genes. ChIP: Peaks. Standard: Popular.

🏠 Known🏠 De novo🏠 Annotation🏠 ChIP-Seq

# HOMER
print("=== Basic Usage ===")
print(""
HOMER:
  - Known + de novo
  - Fast
  - Popular

Command:
  findMotifsGenome.pl peaks.bed hg38 homer_output \
    -size 200 \
    -mask

Options:
  -size: Peak size
  -mask: Repeat masking
  -bg: Background
  -len: Motif lengths
  -S: Number of motifs

Output:
  homer_results.html
  knownResults.txt
  homerMotifs.all.meme
""")

print("\n=== Input Types ===")
print(""
BED:
  findMotifsGenome.pl peaks.bed hg38 output

FASTA:
  findMotifs.pl input.fa output

Positions:
  findMotifsGenome.pl positions.txt hg38 output

Peak file:
  findMotifsGenome.pl peaks.txt hg38 output -size 200
""")

print("\n=== Known Motifs ===")
print(""
Known motifs:
  - HOMER database
  - JASPAR
  - Custom

Command:
  findMotifsGenome.pl peaks.bed hg38 output -known

Output:
  knownResults.txt
  Columns:
    - Motif name
    - Consensus
    - P-value
    - Log P
    - Target %
    - Background %
""")

print("\n=== De Novo ===")
print(""
De novo:
  - Discover from scratch
  - Rich motifs
  - Standard

Command:
  findMotifsGenome.pl peaks.bed hg38 output -size 200

Output:
  homerMotifs.all.meme
  homerResults.html
""")

print("\n=== Annotation ===")
print(""
Annotation:
  - Associate peaks with genes
  - Nearest gene
  - GO

Command:
  annotatePeaks.pl peaks.bed hg38 > annotation.txt

Options:
  -go: GO analysis
  -genome: Genome
  -gtf: Custom annotation

Output:
  annotation.txt
  Columns:
    - Peak
    - Nearest gene
    - Distance
    - Feature
""")

print("\n=== HOMER Databases ===")
print(""
Supported:
  - Human (hg19, hg38)
  - Mouse (mm9, mm10)
  - Rat (rn5, rn6)
  - Zebrafish (danRer10)
  - Drosophila (dm3, dm6)
  - C. elegans (ce10)
  - Yeast (sacCer3)
  - Arabidopsis (tair10)
  - Many more
""")

print("\n=== Custom Databases ===")
print(""
Custom:
  - Custom genome
  - Custom motifs

Prepare:
  loadGenome.pl -org custom -fasta genome.fa -gtf genes.gtf

Use:
  findMotifsGenome.pl peaks.bed custom output
""")

print("\n=== Best Practices ===")
print(""
1. Use -mask
2. Use appropriate size
3. Use background
4. Check known motifs
5. Check de novo
6. Annotate
7. Validate
8. Report
""")

⚡ DREME & STREME

🎯 Fast Short Motif Discovery

DREME and STREME are fast, enumerative tools for discovering short motifs, especially in ChIP-Seq peaks.

DREME & STREME

DREME: Short motifs. STREME: Fast, sensitive. Enumerative: Complete. Discriminative: Primary vs control. Fast: Efficient. ChIP: Peaks.

⚡ DREME⚡ STREME⚡ Fast⚡ Discriminative

# DREME & STREME
print("=== DREME ===")
print(""
DREME:
  - Short motifs
  - Enumerative
  - Fast

Command:
  dreme -p peaks.fa -o dreme_output

Options:
  -p: Primary sequences
  -n: Control sequences
  -e: E-value threshold
  -m: Max motifs
  -k: k-mer size

Output:
  dreme.txt
  dreme.html
""")

print("\n=== STREME ===")
print(""
STREME:
  - Fast
  - Sensitive
  - Enumerative

Command:
  streme --p peaks.fa --o streme_output

Options:
  --p: Primary sequences
  --n: Control sequences
  --thresh: Threshold
  --verbosity: Verbosity

Output:
  streme.txt
  streme.html
""")

print("\n=== Comparison ===")
print(f"{'Feature':<20} {'DREME':<20} {'STREME':<20}")
print("-" * 60)
print(f"{'Speed':<20} {'Fast':<20} {'Very fast':<20}")
print(f"{'Motif length':<20} {'Short (4-8)':<20} {'Short (4-15)':<20}")
print(f"{'Sensitivity':<20} {'Good':<20} {'Excellent':<20}")
print(f"{'Discriminative':<20} {'Yes':<20} {'Yes':<20}")
print(f"{'Output':<20} {'MEME':<20} {'MEME':<20}")

print("\n=== Discriminative ===")
print(""
Discriminative:
  - Primary vs control
  - Enriched motifs
  - Reduces false positives

Example:
  Primary: ChIP-Seq peaks
  Control: Shuffled / background

Output:
  - Motifs enriched in primary
  - Statistically significant
  - Ranked by E-value
""")

print("\n=== Workflow ===")
print(""
1. Extract sequences:
   bedtools getfasta -fi genome.fa -bed peaks.bed -fo peaks.fa

2. Generate control:
   fasta-shuffle-letters peaks.fa control.fa

3. Run DREME:
   dreme -p peaks.fa -n control.fa -o dreme_output

4. Compare:
   tomtom dreme_output/dreme.txt JASPAR2020.meme -oc tomtom
""")

print("\n=== Best Practices ===")
print(""
1. Use control
2. Use appropriate k
3. Check E-value
4. Compare to databases
5. Combine with MEME
6. Validate
7. Report
""")

🔎 Motif Scanning (FIMO, MAST)

🎯 Finding Known Motifs in Sequences

Motif scanning searches sequences for matches to known motifs using PWMs, producing p-values and positions.

Scanning

FIMO: Individual motifs. MAST: Multiple motifs. P-value: Significance. PWM: Scoring. Genome: Scan. Annotation: Sites.

🔎 FIMO🔎 MAST🔎 P-value🔎 Genome scan

# Motif Scanning
print("=== FIMO ===")
print(""
FIMO:
  - Find Individual Motif Occurrences
  - PWM matching
  - P-values

Command:
  fimo motif.meme genome.fa > fimo.txt

Options:
  --thresh: P-value threshold (1e-4)
  --bgfile: Background
  --parse-genomic-coord
  --max-stored-scores

Output:
  fimo.txt
  fimo.html

Columns:
  - Motif ID
  - Sequence name
  - Start
  - Stop
  - Strand
  - Score
  - P-value
  - Q-value
  - Matched sequence
""")

print("\n=== MAST ===")
print(""
MAST:
  - Motif Alignment and Search Tool
  - Multiple motifs
  - E-values

Command:
  mast motif.meme genome.fa > mast.txt

Options:
  -mt: Motif threshold
  -ev: E-value
  -di: Diagram

Output:
  mast.txt
  mast.html
""")

print("\n=== Example FIMO ===")
print(""
#motif_id  motif_alt_id  sequence_name  start  stop  strand  score  p-value  q-value  matched_sequence
p53        MA0001.1      chr1           1000   1010  +       12.5   1e-5     1e-4     GGACATGTCC
p53        MA0001.1      chr1           2000   2010  -       11.8   2e-5     1e-4     GGACATGTCC
CTCF       MA0139.1      chr1           3000   3014  +       15.2   1e-6     1e-5     CCCTCCTGG
""")

print("\n=== Scanning Workflow ===")
print(""
1. Get motif:
   - JASPAR
   - HOMER
   - MEME output

2. Get sequences:
   - Genome
   - Peaks
   - Promoters

3. Scan:
   fimo motif.meme sequences.fa > fimo.txt

4. Filter:
   awk '$7 < 1e-4' fimo.txt > significant.txt

5. Annotate:
   bedtools intersect -a fimo.bed -b genes.bed
""")

print("\n=== Scanning Genomes ===")
print(""
Genome-wide:
  fimo --parse-genomic-coord motif.meme genome.fa > genome_fimo.txt

Large genomes:
  - Split by chromosome
  - Parallel processing
  - Use background

Memory:
  - Use --max-stored-scores
  - Use --thresh
  - Use --bgfile
""")

print("\n=== Scanning Peaks ===")
print(""
Peaks:
  bedtools getfasta -fi genome.fa -bed peaks.bed -fo peaks.fa
  fimo motif.meme peaks.fa > peaks_fimo.txt

Enrichment:
  - Compare to background
  - AME
  - CentriMo
""")

print("\n=== Best Practices ===")
print(""
1. Use appropriate threshold
2. Use background
3. Use genomic coordinates
4. Filter by p-value
5. Annotate
6. Compare to ChIP
7. Validate
8. Report
""")

🗄️ Databases (JASPAR, HOCOMOCO)

🎯 Known Motif Collections

Motif databases provide curated collections of known PWMs for TFs, RBPs, and other regulatory factors.

Databases

JASPAR: Open. HOCOMOCO: Human/mouse. CIS-BP: Large. UniPROBE: PBM. TRANSFAC: Commercial. Jolma: HT-SELEX.

🗄️ JASPAR🗄️ HOCOMOCO🗄️ CIS-BP🗄️ UniPROBE

# Motif Databases
print("=== JASPAR ===")
print(""
JASPAR:
  - Open access
  - Curated
  - Multiple species
  - Standard

Collections:
  - CORE: Verified
  - UNVALIDATED
  - PLANTA
  - INSECTA
  - VERTEBRATA

Formats:
  - PFM (frequencies)
  - PWM (log-odds)
  - MEME
  - TRANSFAC

Download:
  https://jaspar.genereg.net/downloads/

URL:
  https://jaspar.genereg.net/
""")

print("\n=== HOCOMOCO ===")
print(""
HOCOMOCO:
  - Human + mouse
  - High quality
  - Comprehensive

Versions:
  - v11 (current)
  - v12 (2024)

Collections:
  - A: High confidence
  - B: Medium
  - C: Low

Download:
  https://hocomoco11.autosome.org/

URL:
  https://hocomoco11.autosome.org/
""")

print("\n=== CIS-BP ===")
print(""
CIS-BP:
  - Large
  - Many species
  - TF motifs

Features:
  - 1000s of TFs
  - Many species
  - Direct + indirect

URL:
  http://cisbp.ccbr.utoronto.ca/
""")

print("\n=== UniPROBE ===")
print(""
UniPROBE:
  - Protein binding microarray
  - PBM data
  - High throughput

Features:
  - 8-mer scores
  - Many TFs
  - Mouse + human

URL:
  http://thebrain.bwh.harvard.edu/uniprobe/
""")

print("\n=== TRANSFAC ===")
print(""
TRANSFAC:
  - Commercial
  - Comprehensive
  - Curated

Features:
  - TF motifs
  - Regulation
  - Pathways

URL:
  https://gene-regulation.com/
""")

print("\n=== Other Databases ===")
print(""
1. JASPAR
2. HOCOMOCO
3. CIS-BP
4. UniPROBE
5. TRANSFAC
6. Jolma 2013
7. HT-SELEX
8. FactorBook
9. ReMap
10. GTRD
""")

print("\n=== Downloading ===")
print(""
JASPAR:
  wget https://jaspar.genereg.net/download/data/2022/CORE/JASPAR2022_CORE_vertebrates_non-redundant_pfms_meme.txt

HOCOMOCO:
  wget https://hocomoco11.autosome.org/final_bundle/hocomoco11/full/HUMAN/mono/HOCOMOCOv11_full_HUMAN_mono_meme_format.meme

CIS-BP:
  wget http://cisbp.ccbr.utoronto.ca/data/2.00/DataFiles/Mus_musculus_2022_08_23_12-41_pm.zip
""")

print("\n=== Best Practices ===")
print(""
1. Use appropriate database
2. Use non-redundant
3. Filter by quality
4. Compare multiple
5. Check species
6. Use latest version
7. Cite
8. Report
""")

🐍 Python Integration

🎯 Motif Discovery from Python

You can run MEME, HOMER, and FIMO from Python and analyze motifs with Biopython, pandas, and logomaker.

Python

subprocess: Run CLI. Biopython: Motifs. pandas: Results. logomaker: Logos. pybedtools: BED. Pipeline: Automate.

🐍 subprocess🐍 Biopython🐍 pandas🐍 logomaker

# Python Integration
print("=== Run MEME ===")
print(""
import subprocess

def run_meme(fasta_file, output_dir="meme_output",
             nmotifs=10, minw=6, maxw=20):
    \"\"\"Run MEME.\"\"\"
    cmd = [
        'meme', fasta_file,
        '-dna',
        '-oc', output_dir,
        '-nmotifs', str(nmotifs),
        '-minw', str(minw),
        '-maxw', str(maxw),
        '-mod', 'zoops'
    ]
    
    result = subprocess.run(cmd, capture_output=True, text=True)
    
    if result.returncode != 0:
        raise RuntimeError(f"MEME failed: {result.stderr}")
    
    return f"{output_dir}/meme.txt"

# Use
# motifs = run_meme("peaks.fa")
""")

print("\n=== Run HOMER ===")
print(""
def run_homer(bed_file, genome="hg38", output_dir="homer_output",
              size=200, mask=True):
    \"\"\"Run HOMER.\"\"\"
    cmd = [
        'findMotifsGenome.pl',
        bed_file,
        genome,
        output_dir,
        '-size', str(size)
    ]
    
    if mask:
        cmd.append('-mask')
    
    result = subprocess.run(cmd, capture_output=True, text=True)
    
    if result.returncode != 0:
        raise RuntimeError(f"HOMER failed: {result.stderr}")
    
    return output_dir

# Use
# output = run_homer("peaks.bed")
""")

print("\n=== Run FIMO ===")
print(""
def run_fimo(motif_file, sequence_file, output_file="fimo.txt",
             thresh=1e-4):
    \"\"\"Run FIMO.\"\"\"
    cmd = [
        'fimo',
        '--thresh', str(thresh),
        '--parse-genomic-coord',
        motif_file,
        sequence_file
    ]
    
    with open(output_file, 'w') as f:
        result = subprocess.run(cmd, stdout=f, stderr=subprocess.PIPE,
                               text=True)
    
    if result.returncode != 0:
        raise RuntimeError(f"FIMO failed: {result.stderr}")
    
    return output_file

# Use
# hits = run_fimo("motif.meme", "genome.fa")
""")

print("\n=== Parse MEME Output ===")
print(""
import pandas as pd
import re

def parse_meme(meme_file):
    \"\"\"Parse MEME output.\"\"\"
    motifs = []
    with open(meme_file) as f:
        content = f.read()
    
    # Find motifs
    for match in re.finditer(
        r'MOTIF (\S+) .*?letter-probability matrix.*?\n(.*?)\n\n',
        content, re.DOTALL
    ):
        name = match.group(1)
        matrix = match.group(2).strip()
        motifs.append({
            'name': name,
            'matrix': matrix
        })
    
    print(f"Motifs: {len(motifs)}")
    return motifs

# Use
# motifs = parse_meme("meme.txt")
""")

print("\n=== Parse FIMO ===")
print(""
def parse_fimo(fimo_file):
    \"\"\"Parse FIMO output.\"\"\"
    df = pd.read_csv(fimo_file, sep='\\t', comment='#')
    
    print(f"Hits: {len(df)}")
    print(f"Motifs: {df['motif_id'].nunique()}")
    print(f"Sequences: {df['sequence_name'].nunique()}")
    
    # Significant
    sig = df[df['p-value'] < 1e-4]
    print(f"Significant: {len(sig)}")
    
    return df

# Use
# hits = parse_fimo("fimo.txt")
""")

print("\n=== Motif Logo ===")
print(""
import logomaker
import pandas as pd
import matplotlib.pyplot as plt

def plot_logo(pwm_data, output="logo.png"):
    \"\"\"Plot motif logo.\"\"\"
    # Create DataFrame
    df = pd.DataFrame(pwm_data, columns=['A', 'C', 'G', 'T'])
    
    # Create logo
    logo = logomaker.Logo(df)
    logo.style_spines(visible=False)
    logo.style_spines(spines=['left', 'bottom'], visible=True)
    
    plt.tight_layout()
    plt.savefig(output)
    plt.show()

# Use
# plot_logo(pwm_data)
""")

print("\n=== Complete Pipeline ===")
print(""
import subprocess
import pandas as pd
from Bio import SeqIO

def motif_pipeline(peaks_bed, genome_fa, genome="hg38",
                   output_dir="motif_output"):
    \"\"\"Complete motif pipeline.\"\"\"
    import os
    os.makedirs(output_dir, exist_ok=True)
    
    # 1. Extract sequences
    print("Extracting sequences...")
    peaks_fa = f"{output_dir}/peaks.fa"
    subprocess.run([
        'bedtools', 'getfasta',
        '-fi', genome_fa,
        '-bed', peaks_bed,
        '-fo', peaks_fa
    ], check=True)
    
    # 2. Run MEME
    print("Running MEME...")
    meme_txt = run_meme(peaks_fa, f"{output_dir}/meme")
    
    # 3. Run HOMER
    print("Running HOMER...")
    run_homer(peaks_bed, genome, f"{output_dir}/homer")
    
    # 4. Compare to JASPAR
    print("Comparing to JASPAR...")
    subprocess.run([
        'tomtom', meme_txt,
        'JASPAR2020.meme',
        '-oc', f"{output_dir}/tomtom"
    ], check=True)
    
    print("Pipeline complete!")
    return output_dir

# Use
# output = motif_pipeline("peaks.bed", "genome.fa")
""")

🎨 Visualization

🎯 Seeing Motifs

Motif visualization is essential for interpretation and publication. Logos, heatmaps, and genome browser tracks are standard.

Visualization

Logo: Sequence. Heatmap: Motifs. IGV: Genome. Tomtom: Comparison. CentriMo: Distribution. Network: Regulation.

🎨 Logo🎨 Heatmap🎨 IGV🎨 Network

# Visualization
print("=== Sequence Logo ===")
print(""
Logo:
  - Height = information content
  - Letter size = frequency
  - Conserved = tall
  - Variable = short

Tools:
  - WebLogo
  - Skylign
  - ggseqlogo (R)
  - logomaker (Python)

WebLogo:
  https://weblogo.berkeley.edu/

Skylign:
  https://skylign.org/

Python:
  import logomaker
  logo = logomaker.Logo(df)
""")

print("\n=== Heatmap ===")
print(""
Heatmap:
  - Motifs x sequences
  - Scores
  - Clustering

Tools:
  - pheatmap (R)
  - ComplexHeatmap (R)
  - seaborn (Python)
  - matplotlib (Python)

Example:
  import seaborn as sns
  sns.clustermap(scores)
""")

print("\n=== IGV ===")
print(""
IGV (Integrative Genomics Viewer):
  - Genome browser
  - Interactive
  - Standard

Load:
  - BED (peaks)
  - FIMO hits
  - Motif tracks
  - ChIP-Seq

Commands:
  igv.sh -g hg38 peaks.bed fimo.bed

URL:
  https://igv.org/
""")

print("\n=== Tomtom ===")
print(""
Tomtom:
  - Compare motifs
  - Heatmap
  - E-values

Output:
  tomtom.html
  tomtom.tsv

Visualization:
  - Query vs target
  - Alignment
  - E-value
""")

print("\n=== CentriMo ===")
print(""
CentriMo:
  - Central enrichment
  - ChIP-Seq
  - Distribution

Command:
  centrimo peaks.fa motif.meme -oc centrimo_output

Output:
  centrimo.html
  Distribution plot
""")

print("\n=== Network ===")
print(""
Network:
  - TF → target
  - Motif → gene
  - Regulation

Tools:
  - Cytoscape
  - igraph
  - networkx
  - RcisTarget
""")

print("\n=== Genome Browser Tracks ===")
print(""
Tracks:
  - Peaks (BED)
  - Motifs (BED)
  - Conservation
  - ChIP-Seq
  - RNA-Seq

Tools:
  - IGV
  - UCSC
  - JBrowse
""")

print("\n=== Best Practices ===")
print(""
1. Use logos
2. Colorblind-friendly
3. High resolution
4. Vector format
5. Annotate
6. Compare
7. Document
8. Publish
""")

⚖️ Tool Comparison

🎯 Which Tool?

Choosing the right motif tool depends on your goal: discovery, scanning, or comparison.

Comparison

MEME: Discovery. DREME: Short. HOMER: Known + de novo. FIMO: Scanning. TOMTOM: Compare. JASPAR: Database.

⚖️ MEME⚖️ HOMER⚖️ FIMO⚖️ JASPAR

Tool	Type	Speed	Best For
MEME	Discovery	Medium	Comprehensive de novo
DREME	Discovery	Fast	Short motifs
STREME	Discovery	Fast	Short + sensitive
HOMER	Discovery	Fast	Known + de novo, annotation
Gibbs	Discovery	Slow	Bayesian sampling
FIMO	Scanning	Fast	Known motifs in sequences
MAST	Scanning	Fast	Multiple motifs
TOMTOM	Comparison	Fast	Compare to databases
AME	Enrichment	Fast	Motif enrichment
MEME-ChIP	Pipeline	Medium	ChIP-Seq peaks

# Decision Guide
print("=== Choose Tool by Task ===")
print(""
De novo discovery:
  → MEME (comprehensive)
  → DREME (short)
  → HOMER (known + de novo)

Known motif scanning:
  → FIMO (individual)
  → MAST (multiple)

Compare to databases:
  → TOMTOM

Motif enrichment:
  → AME
  → CentriMo

ChIP-Seq pipeline:
  → MEME-ChIP
  → HOMER

Protein motifs:
  → MEME
  → HMMER

RNA motifs:
  → MEME
  → HOMER
""")

print("\n=== Speed Comparison ===")
print(f"{'Tool':<20} {'10 seqs':<15} {'100 seqs':<15} {'1000 seqs':<15}")
print("-" * 65)
print(f"{'DREME':<20} {'1s':<15} {'10s':<15} {'1 min':<15}")
print(f"{'STREME':<20} {'1s':<15} {'5s':<15} {'30s':<15}")
print(f"{'MEME':<20} {'10s':<15} {'5 min':<15} {'1 hour':<15}")
print(f"{'HOMER':<20} {'5s':<15} {'1 min':<15} {'10 min':<15}")
print(f"{'FIMO':<20} {'1s':<15} {'10s':<15} {'2 min':<15}")

print("\n=== When to Use Which ===")
print(""
USE MEME:
  ✓ Comprehensive
  ✓ Long motifs
  ✓ Publication

USE DREME:
  ✓ Short motifs
  ✓ Fast
  ✓ ChIP-Seq

USE HOMER:
  ✓ Known + de novo
  ✓ Annotation
  ✓ ChIP-Seq

USE FIMO:
  ✓ Known motifs
  ✓ Genome scan
  ✓ Variants

USE TOMTOM:
  ✓ Compare motifs
  ✓ Database match

USE MEME-ChIP:
  ✓ ChIP-Seq pipeline
  ✓ Comprehensive
""")

💼 Use Cases

🎯 Where Motif Discovery Is Used

Motif discovery is used in virtually every area of regulatory genomics: ChIP-Seq, ATAC-Seq, variants, and networks.

Use Cases

ChIP-Seq: TF motifs. ATAC-Seq: Regulatory. Enhancers: Discovery. Variants: Impact. Networks: Regulation. Disease: Mechanisms.

💼 ChIP-Seq💼 ATAC-Seq💼 Variants💼 Networks

# Use Cases
print("=== 1. ChIP-Seq Analysis ===")
print(""
Use case:
  - TF binding sites
  - Peaks
  - Motif discovery

Workflow:
  1. ChIP-Seq
  2. Peak calling
  3. Extract sequences
  4. Motif discovery (MEME, HOMER)
  5. Compare to JASPAR
  6. Validate

Examples:
  - p53 motifs
  - CTCF motifs
  - OCT4 motifs
  - NF-κB motifs

Tools:
  - MEME-ChIP
  - HOMER
  - ChIPseeker
""")

print("\n=== 2. ATAC-Seq Analysis ===")
print("""
Use case:
  - Open chromatin
  - Regulatory elements
  - TF motifs

Workflow:
  1. ATAC-Seq
  2. Peak calling
  3. Motif discovery
  4. TF assignment
  5. Networks

Examples:
  - Cell types
  - Development
  - Cancer

Tools:
  - HOMER
  - MEME
  - chromVAR
""")

print("\n=== 3. Enhancer Discovery ===")
print(""
Use case:
  - Distal regulatory
  - H3K27ac peaks
  - TF motifs

Workflow:
  1. H3K27ac ChIP-Seq
  2. Peak calling
  3. Motif discovery
  4. TF assignment
  5. Networks

Examples:
  - Cell-type specific
  - Cancer enhancers
  - Super-enhancers

Tools:
  - HOMER
  - MEME
  - ROSE
""")

print("\n=== 4. Regulatory Variants ===")
print(""
Use case:
  - GWAS hits
  - eQTLs
  - Motif disruption

Workflow:
  1. Variant
  2. Reference motif
  3. Alternate motif
  4. Score change
  5. Impact

Examples:
  - Cancer
  - Autoimmune
  - Neurodegenerative

Tools:
  - FIMO
  - motifbreakR
  - atSNP
  - JASPAR
""")

print("\n=== 5. Gene Regulatory Networks ===")
print(""
Use case:
  - TF → target genes
  - Motif in promoters
  - Network inference

Workflow:
  1. ChIP-Seq / ATAC
  2. Motif discovery
  3. Target genes
  4. Network
  5. Analysis

Tools:
  - HOMER
  - MEME
  - RcisTarget
  - SCENIC
""")

print("\n=== 6. Splicing Regulation ===")
print(""
Use case:
  - Splice sites
  - ESE/ESS
  - ISE/ISS

Workflow:
  1. RNA-Seq
  2. Splicing analysis
  3. Motif discovery
  4. Validation

Tools:
  - MEME
  - SpliceAid
  - ESEfinder
  - MaxEntScan
""")

print("\n=== 7. Protein Motifs ===")
print(""
Use case:
  - Domains
  - Active sites
  - Binding sites

Workflow:
  1. Sequences
  2. Motif discovery
  3. Compare to PROSITE
  4. Validate

Examples:
  - Walker A
  - Zinc finger
  - N-glycosylation

Tools:
  - MEME
  - PROSITE
  - Pfam
""")

print("\n=== 8. Disease Mechanisms ===")
print(""
Use case:
  - Regulatory variants
  - Enhancer hijacking
  - Motif disruption

Workflow:
  1. Variant
  2. Motif
  3. Impact
  4. Mechanism
  5. Treatment

Examples:
  - Cancer
  - Autoimmune
  - Neurodegenerative
""")

print("\n=== Comparison ===")
print(f"{'Use Case':<25} {'Tool':<20} {'Output':<20}")
print("-" * 65)
print(f"{'ChIP-Seq':<25} {'MEME-ChIP':<20} {'Motifs':<20}")
print(f"{'ATAC-Seq':<25} {'HOMER':<20} {'TF motifs':<20}")
print(f"{'Enhancers':<25} {'HOMER':<20} {'TF assignment':<20}")
print(f"{'Variants':<25} {'FIMO':<20} {'Impact':<20}")
print(f"{'Networks':<25} {'SCENIC':<20} {'Network':<20}")
print(f"{'Splicing':<25} {'MaxEntScan':<20} {'Splice sites':<20}")
print(f"{'Protein':<25} {'MEME':<20} {'Domains':<20}")
print(f"{'Disease':<25} {'motifbreakR':<20} {'Mechanism':<20}")

❓ FAQ

🎯 Frequently Asked Questions

Answers to common questions about motif discovery.

What is motif discovery?

What is a Position Weight Matrix (PWM)?

A Position Weight Matrix (PWM), also called a Position-Specific Scoring Matrix (PSSM), is a representation of a motif. It gives a score for each nucleotide (or amino acid) at each position of the motif, based on observed frequencies in a set of aligned binding sites, often with a background model correction.

Which motif discovery tool should I use?

Use MEME for comprehensive de novo discovery, DREME or STREME for fast short motifs, HOMER for combined known + de novo discovery and annotation, and MEME-ChIP for ChIP-Seq peak analysis. For scanning known motifs, use FIMO (MEME suite) or the JASPAR database with tools like TFBSTools.

What is the difference between de novo and known motif discovery?

De novo motif discovery finds motifs from scratch using only the input sequences (e.g., MEME, DREME, HOMER). Known motif discovery scans input sequences for matches to a database of known motifs (e.g., JASPAR, HOCOMOCO, CIS-BP), using tools like FIMO or MAST. They are complementary.

What is a motif logo?

A motif logo (sequence logo) is a graphical representation of a motif where the height of each letter is proportional to its information content (conservation) at that position and the size of each letter reflects its frequency. It's the standard way to visualize PWMs and is produced by WebLogo, Skylign, or ggseqlogo.

How do I choose the right motif length?

Most TF motifs are 6-20 bp. Use MEME's -minw and -maxw options, or HOMER's default range. Shorter motifs (6-8 bp) have higher false positive rates; longer motifs (15-20 bp) are more specific but may be too restrictive. Try multiple lengths and compare.

What is the difference between MEME and DREME?

MEME uses probabilistic (EM) methods to find longer, more complex motifs with PWM output. DREME uses enumerative methods to find short (4-8 bp) motifs quickly with a discriminative approach (primary vs control). DREME is faster; MEME is more comprehensive. Both are part of the MEME Suite.

What is FIMO and when should I use it?

FIMO (Find Individual Motif Occurrences) scans sequences for matches to known PWMs, reporting positions, scores, and p-values. Use it to find where a known motif occurs in a genome or set of peaks, to test variant impact on binding, or to annotate regulatory regions. It's part of the MEME Suite.

How do I find motifs in ChIP-Seq peaks?

Extract peak sequences with bedtools getfasta, then run MEME-ChIP or HOMER. MEME-ChIP runs MEME, DREME, CentriMo, and TOMTOM automatically. HOMER performs known + de novo discovery and annotation. Both handle background and produce comprehensive HTML reports.

How do I interpret motif E-values?

E-value is the expected number of motifs with a given score or better occurring by chance in a database of a given size. Lower is better: E-value < 1e-10 = very significant, 1e-10 to 1e-4 = significant, > 1e-4 = marginal. Always consider E-value, p-value, and biological context together.

# Quick Reference
print("=== Quick Answers ===")
print(""
Q: What is motif discovery?
A: Finding recurring patterns in sequences.

Q: What is PWM?
A: Position Weight Matrix; motif representation.

Q: Best discovery tool?
A: MEME (comprehensive), DREME (fast).

Q: Known + de novo?
A: HOMER.

Q: ChIP-Seq pipeline?
A: MEME-ChIP, HOMER.

Q: Scan known motifs?
A: FIMO, MAST.

Q: Compare motifs?
A: TOMTOM.

Q: Motif databases?
A: JASPAR, HOCOMOCO, CIS-BP.

Q: Motif logo?
A: WebLogo, Skylign, logomaker.

Q: Free tools?
A: Yes — MEME Suite, HOMER, JASPAR.
""")

🚀 Future

🎯 Where Motif Discovery Is Going

The future is deep learning, single-cell, multi-omics, 3D genomics, and clinical.

Future Directions

Deep learning: CNNs, transformers. Single-cell: scATAC. Multi-omics: Integration. 3D: Hi-C, loops. Clinical: Variants. Spatial: Location.

🚀 Deep learning🚀 Single-cell🚀 Multi-omics🚀 3D genomics

# Future of Motif Discovery
print("=== 1. Deep Learning ===")
print(""
Trend: Deep learning

Ideas:
  - CNN for motif discovery
  - Transformer models
  - Attention
  - Learned representations

Examples:
  - DeepBind
  - BPNet
  - ChromBPNet
  - Enformer

Future:
  - End-to-end
  - Multi-modal
  - Variant effect
  - Clinical
""")

print("\n=== 2. Single-Cell ===")
print(""
Trend: Single-cell

Current:
  - Bulk
  - Average

Single-cell:
  - scATAC-Seq
  - scChIP-Seq
  - Per-cell motifs
  - Cell-type specific

Tools:
  - ArchR
  - Signac
  - chromVAR
  - SCENIC

Future:
  - Cell-type specific
  - Rare populations
  - Developmental
""")

print("\n=== 3. Multi-Omics ===")
print(""
Trend: Integrate everything

Current:
  - Single modality
  - Separate analysis

Multi-omics:
  - ChIP + RNA
  - ATAC + RNA
  - Hi-C + ATAC
  - Single-cell multi-omics

Tools:
  - Seurat
  - MOFA
  - ArchR
  - Signac

Future:
  - Unified models
  - Regulatory networks
  - Systems biology
""")

print("\n=== 4. 3D Genomics ===")
print(""
Trend: 3D genome

Current:
  - 1D sequence
  - Linear motifs

3D:
  - Hi-C
  - TADs
  - Loops
  - Enhancer-promoter

Tools:
  - HiCExplorer
  - Juicer
  - Cooler
  - LoopNet

Future:
  - 3D motif context
  - Loop motifs
  - Clinical
""")

print("\n=== 5. Clinical ===")
print(""
Trend: Clinical

Current:
  - Research
  - Specialized

Future:
  - Clinical routine
  - Variant interpretation
  - Drug response
  - Diagnostics

Examples:
  - Cancer variants
  - Rare disease
  - Pharmacogenomics
""")

print("\n=== 6. Spatial ===")
print(""
Trend: Spatial

Current:
  - Dissociated
  - Lose location

Spatial:
  - Tissue context
  - Spatial ATAC
  - Spatial motifs
  - Tumor microenvironment

Platforms:
  - Spatial-ATAC
  - Spatial-CUT&Tag
  - Slide-tags

Future:
  - 3D
  - Tissue atlases
  - Clinical
""")

# TIMELINE
print("\n=== Timeline ===")
print(""
Past:
  1980s: Consensus
  1990s: MEME, Gibbs
  2000s: HOMER, DREME
  2010s: ChIP-Seq era
  2020s: Deep learning

Present:
  2025: Single-cell
  2026: Multi-omics

Future:
  2030: 3D genomics?
  2035: Clinical?
  2040: Predictive?
""")

print("\n=== Final Thoughts ===")
print(""
Motif discovery is:
  - Fundamental to regulatory genomics
  - Needed for ChIP-Seq
  - Needed for ATAC-Seq
  - Needed for variants
  - Rapidly improving

The future is:
  - Deep learning
  - Single-cell
  - Multi-omics
  - 3D genomics
  - Clinical
  - Spatial

"Motifs are the words
 of the regulatory genome."

Keep exploring. Keep building.
""")