AGO-CLIP miRNA Targets - Usage Guide

Overview

Identify direct miRNA-target interactions experimentally. Three approaches: chimeric methods (CLEAR-CLIP, chimeric eCLIP / miR-eCLIP, CLASH) ligate miRNA-target chimeras for direct pairing; standard AGO HITS-CLIP / eCLIP + computational seed-matching (TargetScan, miRDB) for indirect pairing; HEAP (Halo-Ago2 mouse) for in vivo. Chimeric methods are gold-standard for direct identification; standard AGO-CLIP cannot assign miRNAs without computational prediction. A substantial fraction of miRNA-target interactions are non-canonical (3' compensatory) and missed by seed-only methods.

Prerequisites

conda install -c bioconda samtools bedtools
# Hyb pipeline for chimeras
git clone https://github.com/gkudla/hyb
# miR-eCLIP analysis from Yeo lab

Quick Start

Tell your AI agent: - "Identify direct miRNA-target pairs with chimeric eCLIP / miR-eCLIP" - "AGO HITS-CLIP + TargetScan 7mer-m8 + miRNA expression filter for computational pairing" - "miR-eCLIP for hsa-miR-21 deep targets with probe enrichment" - "Run Hyb pipeline on CLEAR-CLIP data to extract chimeras" - "Why does my AGO-CLIP have 50 candidate miRNAs per peak? No miRNA assignment without chimera" - "Cross-reference TargetScan predictions with AGO-CLIP peaks for high-confidence" - "HEAP mouse strain for in vivo - won't work in human"

Example Prompts

Chimeric Methods

"Hyb pipeline with bowtie2 alignment mode for short miRNA chimeras"

"miR-eCLIP for hsa-miR-21 with probe enrichment"

Standard AGO-CLIP + Computational

"Call peaks from AGO eCLIP, then scan with TargetScanHuman 8.0 conserved predictions"

"Filter for 7mer-m8 / 8mer seeds; drop 6mer (too noisy)"

Expression Filter

"Filter miRNA-target predictions by miRNAs > 100 TPM in matched small-RNA-seq"

Non-Canonical Pairing

"Chimeric methods capture 3' compensatory pairing that TargetScan misses"

"RNAhybrid for full miRNA-target duplex prediction"

Diagnostics

"Chimera rate < 1% - enrich with miR-eCLIP probes for specific miRNAs"

"TargetScan predicts thousands per miRNA - require CLIP peak overlap"

"HEAP results don't translate to human - mouse-only method"

What the Agent Will Do

Identify experimental setup: chimeric eCLIP / standard AGO-CLIP / HEAP / CLEAR-CLIP
For chimeric: Hyb pipeline (bowtie2 mode) to extract miRNA-mRNA chimeras
For standard AGO-CLIP: CLIPper / Skipper peaks + TargetScan seed-matching + miRNA expression filter
Apply canonical seed rules: 7mer-m8 / 7mer-A1 / 8mer; report 6mer separately
For non-canonical / 3' compensatory: chimeric method or RNAhybrid full duplex
Cross-validate: chimeric direct + TargetScan computational + matched small-RNA-seq
Report per-miRNA target counts; flag rare miRNAs (< 100 TPM)
For specific miRNA deep profiling: miR-eCLIP probe enrichment

Tips

Chimeric methods are gold standard for direct pairing. Other methods need computational inference.
miR-eCLIP enriches specific miRNAs via probe capture. For deep targets of one or a few miRNAs.
Hyb pipeline needs bowtie2 mode for short miRNAs. BLAST is too stringent.
Standard chimera rate is 1-5%. Sequence deep (200M+) for global chimera profiling.
TargetScan + CLIP overlap is high-confidence. TargetScan alone over-predicts.
Filter by miRNA expression > 100 TPM. Many miRNAs in databases not expressed in your cell type.
6mer matches dominate but are weak. Report separately; require 7mer-m8 / 8mer for high confidence.
3' compensatory misses in seed-only methods. A substantial fraction of miRNA-target interactions are non-canonical.
HEAP is mouse-only. Use eCLIP / chimeric eCLIP for human.

clip-seq/clip-peak-calling - AGO CLIP peaks
clip-seq/binding-site-annotation - 3' UTR annotation
clip-seq/clip-motif-analysis - Seed motif scan
clip-seq/differential-clip - miRNA perturbation
clip-seq/m6a-clip - DART-seq APOBEC1 fusion
small-rna-seq/target-prediction - TargetScan / miRDB
small-rna-seq/differential-mirna - miRNA expression
small-rna-seq/mirdeep2-analysis - miRNA discovery