Predict RBP binding from RNA sequence using deep learning models (RBPNet sequence-to-signal, RNAProt RNN, GraphProt2 GCN with structure, DeepCLIP, DeepRiPe multi-modal CNN) for variant-effect prediction, in silico binding-site discovery, model interpretation, and transfer learning from CLIP and RBNS datasets. Use when computational prediction of RBP binding from sequence is needed, evaluating variant effects on binding without further wet-lab experiments, comparing model performance, or training a custom model on ENCODE eCLIP data.
Standardmäßig ist der Prompt ausgewählt, der zuerst die Quelle prüft. Sie können zu einem direkten Befehl wechseln oder eine lokale Kopie herunterladen.
Quelldateien prüfen
Lesen Sie SKILL.md und alle von SkillsMP angezeigten Begleitdateien, bevor Sie sich für eine Installation entscheiden.
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Ein direkter Befehl überspringt den Prüf-Prompt. Prüfen Sie die Quelle, bevor Sie ihn ausführen.
Predict RBP binding from RNA sequence using deep learning models (RBPNet sequence-to-signal, RNAProt RNN, GraphProt2 GCN with structure, DeepCLIP, DeepRiPe multi-modal CNN) for variant-effect prediction, in silico binding-site discovery, model interpretation, and transfer learning from CLIP and RBNS datasets. Use when computational prediction of RBP binding from sequence is needed, evaluating variant effects on binding without further wet-lab experiments, comparing model performance, or training a custom model on ENCODE eCLIP data.
tool_type
python
primary_tool
RBPNet
Version Compatibility
Reference examples tested with: RBPNet (Horlacher et al 2023 github), RNAProt 0.5+, GraphProt2 (Uhl et al 2021 github), DeepCLIP 1.0+ (Gronning 2020), DeepRiPe (Ohler lab), pytorch 2.2+, tensorflow 2.15+, scikit-learn 1.4+, biopython 1.83+, transformers 4.40+ (for RNA foundation models).
Before using code patterns, verify installed versions match. If versions differ:
Python: pip show <package> then help(module.function) to check signatures
Frameworks: check pytorch / tensorflow versions; reproducibility depends on framework version
If code throws ImportError, AttributeError, or TypeError, introspect the installed package and adapt the example to match the actual API rather than retrying.
CLIP-seq Deep Learning
"Predict RBP binding from RNA sequence using deep learning" -> Train or apply neural networks that learn the sequence (and optionally structure) preference of an RBP from CLIP-seq peaks or single-nucleotide crosslink sites. The output is per-base or per-site binding probability for any input sequence, enabling: (a) variant-effect prediction at heterozygous SNPs; (b) in silico binding-site discovery on transcripts not covered by CLIP; (c) systematic comparison across RBPs via shared model architectures; (d) interpretation via attribution / saliency to recover RBP-specific motifs and structural preferences. Modern models (RBPNet 2023) predict per-nucleotide crosslink count distributions rather than binary peak/non-peak, providing single-nt resolution outputs.
Python (RBPNet sequence-to-CL signal): import rbpnet; model = rbpnet.load_pretrained('RBP_name'); predictions = model.predict(sequence) produces per-base CL count distribution
The benchmarks (RNAProt paper, 2021): RNAProt AUC 87-89%; DeepCLIP 84-87%; GraphProt 82-84%. RBPNet (2023) is the modern sequence-to-signal model that predicts per-nt CL distributions at single-nucleotide resolution rather than binary site classification.
Models Taxonomy
Model
Architecture
Input
Output
Resolution
Strength
Fails when
RBPNet (Horlacher et al 2023)
Sequence-to-signal CNN
Sequence
Per-nt CL count distribution
Single-nt
Modern single-nt resolution; predicts CL distribution not binary
New (2023); fewer pretrained RBPs
RNAProt (Uhl 2021)
GRU RNN
Sequence
Binary binding probability
Site/peak
Highest AUC in benchmark (87-89%); feature-rich (RNAplfold structure)
Single-prediction; not per-base profile
GraphProt (Maticzka 2014)
Graph kernel + SVM
Sequence + structure
Binary
Site/peak
Original structure-aware; well-validated
Older; superseded by GraphProt2
GraphProt2 (Uhl 2021)
Graph Convolutional Network (GCN)
Variable-length sequence + structure
Nucleotide-wise binding profile
Per-nt
Variable-length input; structure-aware
Slow to train; needs GPU
DeepCLIP (Gronning 2020)
CNN + BiLSTM
Sequence
Binary
Site
Fast; good benchmarks
Sequence-only; no structure
DeepRiPe (Ghanbari 2020)
Multi-modal CNN
Sequence + region type
Binary
Site
Multi-input; good ENCODE benchmark
Sequence/region must be pre-extracted
iDeep / iDeepE (Pan 2018)
CNN ensemble
Sequence
Binary
Site
Ensemble approach
Older; few pretrained models
Pysster (Budach 2018)
CNN-LSTM
Sequence
Binary
Site
Generic framework
Less RBP-specific
DeepBind (Alipanahi 2015)
CNN
Sequence
Binary
Site
First deep-learning RBP model
Outdated; superseded
Basenji-style multi-task CNN
Multi-task CNN
Sequence
Per-task profile
Per-nt
Joint learning across RBPs
Computational overhead; no dedicated CLIP tool
RNA foundation models (RNAErnie, RNA-FM)
Transformer pretrained on RNA
Sequence
Embeddings (downstream task)
Embedding
Transfer learning across RBPs
Foundation model trained at depth; fine-tuning needed
Methodology evolves; verify the latest publication on RBP deep learning (RBPNet 2023 is the current state-of-the-art per-nt resolution model). RNA foundation models (RNA-FM, RNAErnie) are emerging in 2024 as transfer-learning backbones; fine-tuning on CLIP data for specific RBPs is the next-generation approach.
Critical Choice: Binary Classification vs Sequence-to-Signal
Binary classification (RNAProt, DeepCLIP, DeepRiPe, GraphProt2): Train on labeled site vs background; predict probability of binding for an input sequence. Output: per-sequence score. Pro: simple framework; mature benchmarks. Con: discards single-nt CL distribution information; binary decision boundary.
Sequence-to-signal (RBPNet): Train on per-nt CL count distributions from PureCLIP or CTK CITS output; predict per-base CL count for input sequence. Output: per-nt profile. Pro: single-nt resolution; preserves CL count information; matches biology (CL is a sharp signal, not a region). Con: newer (2023); fewer pretrained models; harder to interpret with classical motif tools.
Goal
Model
Predict RBP binding probability for an input sequence
RNAProt or DeepRiPe
Predict per-base CL distribution
RBPNet
Variant-effect at heterozygous SNP
RBPNet or DeepRiPe (per-base output)
Compare RBP preferences across ENCODE
Multi-task model (Basenji-style)
Transfer learning across RBPs
RNA foundation model + fine-tune
In silico screening of variants
RBPNet (per-base) for genome-wide
Motif interpretation via attribution
DeepRiPe or GraphProt2 (interpretable)
Custom training on new CLIP data
RNAProt (easiest pipeline)
Production-grade per-base prediction
RBPNet 2023
Variant-Effect Prediction Workflow
Apply a pretrained or custom-trained model to predict the change in binding upon a sequence variant.
import torch
from rbpnet import RBPNet # hypothetical API; verify per-package documentation# Load pretrained model for specific RBP
model = RBPNet.load_pretrained('TARDBP_HEK293T')
# Reference and alternative sequences around a variant
ref_seq = 'CTGTACTGCAGTAGCATGCTAGCATGCTAGCAT'# 32 nt window centered on variant
alt_seq = 'CTGTACTGCAGTAGCATGCTAGCATGCTAGCAA'# Variant: T -> A at position 32# Predict per-base CL distribution for both
ref_pred = model.predict(ref_seq) # shape: (32, 1) - per-base CL probability
alt_pred = model.predict(alt_seq)
# Variant effect: log2 fold change in summed binding signalimport numpy as np
effect = np.log2((alt_pred.sum() + 1e-9) / (ref_pred.sum() + 1e-9))
print(f'Variant effect (log2 FC): {effect:.4f}')
# Strong-effect variant: |log2 FC| > 1.0# Mid-effect: 0.5 - 1.0# Weak: < 0.5
For genome-wide variant scoring: apply this in batch to all GWAS variants overlapping the RBP's binding regions. The output is a per-variant log2 FC; downstream Mendelian randomization or fine-mapping integrates with phenotype-association statistics.
Training a Custom Model (RNAProt Example)
RNAProt is the most accessible training framework. It accepts peak BED + background BED + genome FASTA.
Goal: Train a chromosome-split RNN classifier from CLIP peaks to predict RBP binding probability on arbitrary input sequences, with held-out evaluation on a chromosome-distinct test set.
Approach: Generate a GC-matched 3' UTR background, split peaks and background by chromosome (train chr1-20, test chr21-22) to prevent gene-neighbor leakage, train RNAProt for 50 epochs at batch size 64, and evaluate held-out AUC against the ENCODE benchmark target of 0.85-0.89.
Model false positive; or transient binding not captured by CLIP
Check RBNS prediction; in vitro Kd
Model predicts low; CLIP has strong peak
Model false negative; or non-canonical / context-dependent
Investigate training data; structure-dependent?
Model AUC 0.95 in test; fails on novel sequences
Train/test leakage; chromosome split needed
Re-train with proper split
Variant effect log2 FC inconsistent across windows
Window-size sensitivity
Use model native window; average across shifts
Pretrained for related RBP works on target
Cross-RBP transfer
Transfer learning may work for paralogs
GraphProt2 underperforms sequence-only
Structure ensemble not provided
Fix structure input
Saliency motif noisy
Vanilla gradient method
Use DeepLIFT or TF-MoDISco
Train/test split by random outperforms chromosome split
Leakage from gene-neighbor sequences
Trust chromosome-split AUC
Operational rule for high-confidence variant-effect: (a) Use RBPNet per-base output; (b) compute log2 FC at variant position summed over 50 nt window; (c) cross-validate with another model (DeepRiPe); (d) cross-reference with overlapping eCLIP peak if available; (e) report |log2 FC| > 1 as strong effect.
Common Errors
Error / symptom
Cause
Solution
AUC 0.5 on test
Train data leak or random shuffle
Verify chromosome split
Model trained on 1 epoch
Default optimizer state
Train 30-50 epochs with validation
GPU OOM
Batch size too large
Reduce batch size to 32
Variant effect log2 FC very large (> 10)
Reference sequence not in training distribution
Verify input sequence reasonable
Pretrained model not found
RBP not in pretrained list
Train custom; or use closest paralog
Structure flag without input
GraphProt2 misuse
Pre-compute RNAfold structure
Saliency map flat
Model architecture too shallow
Use DeepRiPe / RBPNet (deeper)
Per-class accuracy uneven
Class imbalance
Use balanced sampling or class weights
Cross-RBP transfer fails
RBPs unrelated
Limit transfer to paralogs or use foundation model
Custom training crashes
RAM / GPU exhausted
Smaller batch; cache embeddings
References
Alipanahi B et al 2015 Nat Biotechnol 33:831 (DeepBind, first DL RBP model)
Maticzka D et al 2014 Genome Biol 15:R17 (GraphProt with structure)
Uhl M et al 2021 bioRxiv 850024 (GraphProt2 with GCN; preprint)
Gronning AGB et al 2020 Nucleic Acids Res 48:7099 (DeepCLIP)
Ghanbari M, Ohler U 2020 Genome Res 30:214 (DeepRiPe multi-modal)
Pan X, Shen HB 2018 Bioinformatics 34:3427 (iDeepE)
Budach S, Marsico A 2018 Bioinformatics 34:3035 (Pysster)
Uhl M et al 2021 GigaScience 10:giab054 (RNAProt RNN)
Horlacher M, Wagner N, Moyon L et al 2023 Genome Biol 24:180 (RBPNet sequence-to-signal at single-nt)
Shrikumar A et al 2018 arXiv (TF-MoDISco interpretation)
Chen J et al 2022 arXiv:2204.00300 (RNA-FM foundation model, preprint)
Wang N et al 2024 Nat Mach Intell 6:548 (RNAErnie)
Related Skills
clip-seq/clip-motif-analysis - Motif analysis is the classical alternative
clip-seq/crosslink-site-detection - Single-nt CL sites for sequence-to-signal models
clip-seq/clip-peak-calling - Peak BEDs for binary classification training
