[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M73B8PTYNKW8GP3SCCHKABKD":3,"skill-same:01M73B8PTYNKW8GP3SCCHKABKD":48},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M73B8PTYNKW8GP3SCCHKABKD","arboreto-grn-inference",null,"GRN inference from expression via GRNBoost2 (gradient boosting) or GENIE3 (Random Forest). Load matrix, filter by TFs, infer TF-target-importance links, save network. Dask-parallelized to single-cell scale. Core SCENIC component.","bg-szy\u002FTOP-SKILLS","bg-szy","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F106576190?v=4","01M73A4E8KPV67SFXRY8XZGNBX","ShuZhiyong","01M73A4E96MQ294MHM74TZBVQD",6,"BSD-3-Clause","https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS\u002Fblob\u002Fmaster\u002Fskills\u002Fawesome-skills\u002Farboreto-grn-inference\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: \"arboreto-grn-inference\"\ndescription: \"GRN inference from expression via GRNBoost2 (gradient boosting) or GENIE3 (Random Forest). Load matrix, filter by TFs, infer TF-target-importance links, save network. Dask-parallelized to single-cell scale. Core SCENIC component.\"\nlicense: \"BSD-3-Clause\"\n---\n\n# Arboreto GRN Inference\n\n## Overview\n\nArboreto infers gene regulatory networks (GRNs) from gene expression data using parallelized tree-based regression. For each target gene, it trains a regression model with all other genes (or a specified TF list) as features and emits TF-target-importance triplets. It provides two interchangeable algorithms -- GRNBoost2 (gradient boosting, fast) and GENIE3 (Random Forest, classic) -- sharing identical input\u002Foutput formats. Computation is Dask-parallelized, scaling from laptop cores to HPC clusters.\n\n## When to Use\n\n- Inferring transcription factor-to-target gene regulatory relationships from bulk RNA-seq expression data\n- Building gene regulatory networks from single-cell RNA-seq count matrices (cells as rows, genes as columns)\n- Generating the adjacency matrix (Step 1) of the pySCENIC regulatory analysis pipeline\n- Comparing regulatory network structure across experimental conditions (e.g., control vs treatment)\n- Producing consensus regulatory networks by running inference across multiple random seeds\n- Validating GRN results by comparing GRNBoost2 and GENIE3 outputs on the same dataset\n- For downstream regulon identification and activity scoring, use arboreto output with pySCENIC\n- For single-cell preprocessing (QC, normalization, clustering) before GRN inference, use **scanpy-scrna-seq**\n\n## Prerequisites\n\n- **Python packages**: `arboreto`, `pandas`, `numpy`, `dask`, `distributed`, `scikit-learn`, `scipy`\n- **Data requirements**: Gene expression matrix (genes as columns, observations as rows) in TSV\u002FCSV; optionally a TF list file (one gene name per line)\n- **Environment**: Python 3.8+; optional `networkx`, `matplotlib` for visualization\n\n```bash\npip install arboreto distributed networkx matplotlib\n```\n\n## Pre-flight Interview\n\nSettle these with the user before writing any analysis code.\n\n```yaml\ndecisions:\n  - id: D1\n    param: algorithm\n    kind: required\n    source: user\n    ask: \"Infer the network with gradient boosting (faster) or random forests (the original GENIE3 formulation)?\"\n    default: \"GRNBoost2\"\n\n  - id: D2\n    param: expressionMatrix\n    kind: required\n    source: upstream\n    ask: \"Which processed expression matrix, and which cells or samples within it, should the network be inferred from?\"\n    default: null\n\n  - id: D3\n    param: regulatorList\n    kind: required\n    source: literature\n    ask: \"Restrict candidate regulators to a curated transcription-factor list, or let every gene be a candidate regulator?\"\n    default: \"a species-matched TF list\"\n\n  - id: D4\n    param: edgeThreshold\n    kind: required\n    source: user\n    ask: \"The output ranks every regulator-target pair - how many of the top edges should be kept as the network?\"\n    default: null\n\n  - id: D5\n    param: seed\n    kind: never_ask\n    source: data\n    reason: \"Fixes the draw for reproducibility; does not change what the data supports\"\n    default: 0\n\n  - id: D6\n    param: daskClient\n    kind: never_ask\n    source: data\n    reason: \"Execution backend; affects runtime only\"\n    default: \"local\"\n```\n\nD3 is the decision that separates a usable network from a hairball. Leaving\nregulators unrestricted lets any correlated gene appear as a regulator, and the\nresult still returns a fully populated ranked table. D4 has no safe default:\nthe algorithm scores all pairs, so where the list is cut is the user's call\nabout the network they intend to interpret.\n\n## Quick Start\n\nComplete GRN inference in a single block. The `if __name__ == '__main__':` guard is required because Dask spawns worker processes via multiprocessing.\n\n```python\nimport pandas as pd\nfrom arboreto.algo import grnboost2\nfrom arboreto.utils import load_tf_names\n\nif __name__ == '__main__':\n    # Load expression matrix (observations x genes)\n    expression_matrix = pd.read_csv('expression_data.tsv', sep='\\t')\n    tf_names = load_tf_names('tf_list.txt')  # optional TF filter\n\n    # Infer GRN (uses all local cores by default)\n    network = grnboost2(expression_data=expression_matrix,\n                        tf_names=tf_names, seed=777)\n\n    # Filter top links and save\n    top_network = network[network['importance'] > 1.0]\n    top_network.to_csv('grn_output.tsv', sep='\\t', index=False, header=False)\n    print(f\"Inferred {len(network)} links, kept {len(top_network)} above threshold\")\n    # Example: Inferred 185432 links, kept 12876 above threshold\n```\n\n## Workflow\n\n### Step 1: Load Expression Data\n\nArboreto accepts a pandas DataFrame (recommended) or NumPy array. Rows are observations (cells or samples), columns are genes. Gene names must be column headers.\n\n```python\nimport pandas as pd\n\n# From TSV (genes as columns, observations as rows)\nexpression_matrix = pd.read_csv('expression_data.tsv', sep='\\t')\nprint(f\"Shape: {expression_matrix.shape}  \"\n      f\"({expression_matrix.shape[0]} observations x {expression_matrix.shape[1]} genes)\")\n# Example: Shape: (5000, 18654)  (5000 observations x 18654 genes)\n\n# From AnnData (e.g., after scanpy preprocessing)\nimport anndata as ad\nadata = ad.read_h5ad('preprocessed.h5ad')\nexpression_matrix = pd.DataFrame(\n    adata.X.toarray() if hasattr(adata.X, 'toarray') else adata.X,\n    columns=adata.var_names.tolist()\n)\nprint(f\"Converted AnnData: {expression_matrix.shape}\")\n# Example: Converted AnnData: (5000, 18654)\n```\n\n### Step 2: Load Transcription Factor List\n\nProviding a TF list restricts regulators to known transcription factors, reducing computation time and improving biological relevance. If omitted, all genes are treated as potential regulators.\n\n```python\nfrom arboreto.utils import load_tf_names\n\n# From file (one TF name per line)\ntf_names = load_tf_names('human_tfs.txt')\nprint(f\"Loaded {len(tf_names)} transcription factors\")\n# Example: Loaded 1639 transcription factors\n\n# Or define directly\ntf_names = ['MYC', 'TP53', 'SOX2', 'NANOG', 'POU5F1']\n\n# Verify TFs exist in expression matrix columns\ntf_in_data = [tf for tf in tf_names if tf in expression_matrix.columns]\nprint(f\"TFs found in expression data: {len(tf_in_data)}\u002F{len(tf_names)}\")\n# Example: TFs found in expression data: 1583\u002F1639\n```\n\n### Step 3: Configure Dask Client (Optional)\n\nBy default arboreto creates an internal Dask client using all local cores. Create an explicit client for resource control, monitoring, or cluster deployment.\n\n```python\nfrom distributed import LocalCluster, Client\n\n# Custom local client with resource limits\nlocal_cluster = LocalCluster(\n    n_workers=8,\n    threads_per_worker=1,   # avoid GIL contention in scikit-learn\n    memory_limit='4GB'       # per worker\n)\nclient = Client(local_cluster)\nprint(f\"Dashboard: {client.dashboard_link}\")\n# Example: Dashboard: http:\u002F\u002F127.0.0.1:8787\u002Fstatus\n```\n\n### Step 4: Run GRN Inference\n\nCall `grnboost2()` (recommended) or `genie3()`. Both share the same signature and output format.\n\n```python\nfrom arboreto.algo import grnboost2\n\nif __name__ == '__main__':\n    network = grnboost2(\n        expression_data=expression_matrix,\n        tf_names=tf_names,         # 'all' if no TF list\n        client_or_address=client,  # omit to use default local scheduler\n        seed=777,                  # for reproducibility\n        verbose=True               # print progress\n    )\n    print(f\"Inferred {len(network)} regulatory links\")\n    print(network.head())\n    # Example:\n    #       TF  target  importance\n    # 0   MYC   CDK4       3.214\n    # 1   MYC   CCND1      2.871\n    # 2  TP53  CDKN1A      2.654\n```\n\n### Step 5: Filter Results\n\nRaw output contains links for every TF-target pair with non-zero importance. Filter to retain high-confidence regulatory interactions.\n\n```python\n# Strategy 1: Importance threshold\nthreshold = 1.0\nfiltered = network[network['importance'] > threshold]\nprint(f\"Threshold {threshold}: {len(filtered)} links \"\n      f\"({len(filtered)\u002Flen(network)*100:.1f}% retained)\")\n# Example: Threshold 1.0: 12876 links (6.9% retained)\n\n# Strategy 2: Top N links per target gene\ntop_n = 10\ntop_per_target = (network.groupby('target')\n                  .apply(lambda g: g.nlargest(top_n, 'importance'))\n                  .reset_index(drop=True))\nprint(f\"Top {top_n} per target: {len(top_per_target)} links\")\n# Example: Top 10 per target: 186540 links\n```\n\n### Step 6: Save Network\n\nSave as a TSV file with three columns: TF, target, importance.\n\n```python\n# Save full network\nnetwork.to_csv('full_network.tsv', sep='\\t', index=False)\nprint(f\"Saved full network: {len(network)} links\")\n\n# Save filtered network (without header for pySCENIC compatibility)\nfiltered.to_csv('filtered_network.tsv', sep='\\t', index=False, header=False)\nprint(f\"Saved filtered network: {len(filtered)} links\")\n\n# Clean up Dask client if explicitly created\nclient.close()\nlocal_cluster.close()\n```\n\n### Step 7: Visualize Results (Optional)\n\nPlot the top regulatory interactions as a directed network graph.\n\n```python\nimport networkx as nx\nimport matplotlib.pyplot as plt\n\n# Build directed graph from top links\ntop_links = network.nlargest(50, 'importance')\nG = nx.from_pandas_edgelist(\n    top_links, source='TF', target='target',\n    edge_attr='importance', create_using=nx.DiGraph()\n)\n\n# Draw network\nfig, ax = plt.subplots(figsize=(12, 10))\npos = nx.spring_layout(G, k=2, seed=42)\nnx.draw_networkx_nodes(G, pos, node_size=300, node_color='lightblue', ax=ax)\nnx.draw_networkx_labels(G, pos, font_size=7, ax=ax)\nedges = nx.draw_networkx_edges(\n    G, pos, edge_color=[G[u][v]['importance'] for u, v in G.edges()],\n    edge_cmap=plt.cm.Reds, width=1.5, arrows=True, ax=ax\n)\nplt.colorbar(edges, ax=ax, label='Importance')\nax.set_title('Top 50 Regulatory Interactions')\nplt.tight_layout()\nplt.savefig('grn_network.png', dpi=300, bbox_inches='tight')\nprint(\"Saved grn_network.png\")\n```\n\n## Key Parameters\n\n| Parameter | Default | Range \u002F Options | Effect |\n|-----------|---------|-----------------|--------|\n| `expression_data` | (required) | DataFrame or ndarray | Expression matrix, observations x genes |\n| `tf_names` | `'all'` | list of str or `'all'` | Restrict regulators to known TFs; `'all'` uses every gene |\n| `gene_names` | `None` | list of str | Required when `expression_data` is a NumPy array |\n| `client_or_address` | `'local'` | Client, str, or `'local'` | Dask client instance or scheduler address |\n| `seed` | `None` | int | Random seed for reproducibility; always set for replicable results |\n| `verbose` | `False` | bool | Print progress messages during inference |\n\n## Key Concepts\n\n### Algorithm Selection Guide\n\nBoth algorithms follow the same multiple-regression strategy (train one model per target gene, extract feature importances) but differ in the underlying regressor.\n\n| Feature | GRNBoost2 | GENIE3 |\n|---------|-----------|--------|\n| **Method** | Stochastic gradient boosting with early stopping | Random Forest (or ExtraTrees) |\n| **Speed** | Fast -- optimized for large datasets | Slower -- higher per-model cost |\n| **Memory** | Lower (early stopping limits tree depth) | Higher (full forests per target) |\n| **Best for** | Default choice; 10k+ observations, single-cell scale | Comparison with published GENIE3 results; validation |\n| **Import** | `from arboreto.algo import grnboost2` | `from arboreto.algo import genie3` |\n| **Output format** | TF, target, importance (identical) | TF, target, importance (identical) |\n\n**Decision rule**: Start with GRNBoost2. Use GENIE3 only when reproducing published GENIE3 analyses or as an independent validation of GRNBoost2 results.\n\n### Output Format\n\nThe network DataFrame has three columns, sorted by descending importance:\n\n| Column | Type | Description |\n|--------|------|-------------|\n| `TF` | str | Transcription factor (regulator) gene name |\n| `target` | str | Target gene name |\n| `importance` | float | Regulatory importance score (higher = stronger predicted regulation) |\n\nImportance scores are not p-values. They represent feature importance from the tree-based model. Filter by absolute threshold or top-N per target for downstream analysis.\n\n### The `if __name__ == '__main__':` Requirement\n\nDask uses Python's `multiprocessing` module to spawn worker processes. Without the main guard, each spawned process re-executes the script, leading to infinite process spawning. Always wrap arboreto calls in `if __name__ == '__main__':` when running as a script. This is not needed inside Jupyter notebooks.\n\n## Common Recipes\n\n### Recipe: Distributed Computing on a Cluster\n\nWhen to use: datasets too large for a single machine (e.g., 50k+ cells, 20k+ genes). Requires a Dask distributed scheduler running on the cluster.\n\n```bash\n# On head node: start scheduler\ndask-scheduler\n# Output: Scheduler at tcp:\u002F\u002F10.118.224.134:8786\n\n# On each compute node: start workers\ndask-worker tcp:\u002F\u002F10.118.224.134:8786 --nprocs 4 --nthreads 1 --memory-limit 16GB\n```\n\n```python\nfrom distributed import Client\nfrom arboreto.algo import grnboost2\nimport pandas as pd\n\nif __name__ == '__main__':\n    client = Client('tcp:\u002F\u002F10.118.224.134:8786')\n    print(f\"Connected to cluster: {client.scheduler_info()['workers'].__len__()} workers\")\n    # Example: Connected to cluster: 16 workers\n\n    expression_data = pd.read_csv('large_scrna.tsv', sep='\\t')\n    tf_names = pd.read_csv('tf_list.txt', header=None)[0].tolist()\n\n    network = grnboost2(\n        expression_data=expression_data,\n        tf_names=tf_names,\n        client_or_address=client,\n        seed=42, verbose=True\n    )\n    network.to_csv('cluster_grn.tsv', sep='\\t', index=False)\n    print(f\"Cluster inference complete: {len(network)} links\")\n    client.close()\n```\n\n### Recipe: Consensus Network from Multiple Seeds\n\nWhen to use: improve robustness by aggregating networks from multiple random seeds. Links appearing consistently across runs are more reliable.\n\n```python\nimport pandas as pd\nfrom distributed import LocalCluster, Client\nfrom arboreto.algo import grnboost2\n\nif __name__ == '__main__':\n    client = Client(LocalCluster(n_workers=8, threads_per_worker=1))\n    expression_data = pd.read_csv('expression_data.tsv', sep='\\t')\n    tf_names = pd.read_csv('tf_list.txt', header=None)[0].tolist()\n\n    seeds = [42, 123, 456, 789, 1001]\n    all_networks = []\n\n    for seed in seeds:\n        net = grnboost2(expression_data=expression_data,\n                        tf_names=tf_names,\n                        client_or_address=client, seed=seed)\n        net['seed'] = seed\n        all_networks.append(net)\n        print(f\"Seed {seed}: {len(net)} links\")\n\n    combined = pd.concat(all_networks, ignore_index=True)\n\n    # Consensus: mean importance across seeds, keep links found in 3+ runs\n    consensus = (combined.groupby(['TF', 'target'])\n                 .agg(mean_importance=('importance', 'mean'),\n                      n_seeds=('seed', 'nunique'))\n                 .reset_index())\n    consensus = consensus[consensus['n_seeds'] >= 3]\n    consensus = consensus.sort_values('mean_importance', ascending=False)\n    consensus.to_csv('consensus_network.tsv', sep='\\t', index=False)\n    print(f\"Consensus network: {len(consensus)} links (found in 3+ of {len(seeds)} runs)\")\n    # Example: Consensus network: 28453 links (found in 3+ of 5 runs)\n\n    client.close()\n```\n\n### Recipe: pySCENIC Integration\n\nWhen to use: downstream regulatory analysis after arboreto GRN inference. Arboreto produces the adjacency matrix (Step 1 of SCENIC); pySCENIC performs regulon identification (Step 2) and activity scoring (Step 3).\n\n```python\nfrom arboreto.algo import grnboost2\nfrom arboreto.utils import load_tf_names\nimport pandas as pd\n\nif __name__ == '__main__':\n    # Step 1 (arboreto): GRN inference\n    expression_matrix = pd.read_csv('scrna_expression.tsv', sep='\\t')\n    tf_names = load_tf_names('allTFs_hg38.txt')\n\n    adjacencies = grnboost2(\n        expression_data=expression_matrix,\n        tf_names=tf_names, seed=777\n    )\n    adjacencies.to_csv('adjacencies.tsv', sep='\\t', index=False, header=False)\n    print(f\"SCENIC Step 1 complete: {len(adjacencies)} adjacencies\")\n    # Example: SCENIC Step 1 complete: 234567 adjacencies\n\n    # Steps 2-3 use pySCENIC CLI (requires pyscenic installation):\n    # pyscenic ctx adjacencies.tsv hg38_ranking.feather \\\n    #     --annotations_fname motifs-v10nr.hgnc-m0.001-o0.0.tbl \\\n    #     --output regulons.csv\n    #\n    # pyscenic aucell scrna_expression.tsv regulons.csv \\\n    #     --output auc_matrix.csv\n    print(\"Next: run pyscenic ctx (Step 2) and pyscenic aucell (Step 3)\")\n```\n\n## Expected Outputs\n\n- `full_network.tsv` -- complete TF-target-importance table; columns: TF, target, importance\n- `filtered_network.tsv` -- high-confidence subset after threshold or top-N filtering\n- `consensus_network.tsv` -- aggregated network from multi-seed runs; columns: TF, target, mean_importance, n_seeds\n- `adjacencies.tsv` -- arboreto output formatted for pySCENIC input (no header)\n- `grn_network.png` -- directed graph visualization of top regulatory interactions\n\n## Troubleshooting\n\n| Problem | Cause | Solution |\n|---------|-------|----------|\n| `RuntimeError: freeze_support()` or infinite process spawning | Missing `if __name__ == '__main__':` guard | Wrap all arboreto calls in main guard; not needed in Jupyter |\n| `MemoryError` during inference | Expression matrix too large for available RAM | Filter low-variance genes first, reduce to top 5-10k genes, or use distributed cluster |\n| Empty or near-empty network | TF names do not match expression matrix column names | Verify TF names overlap with `expression_matrix.columns`; check case sensitivity |\n| Very slow inference | Using GENIE3 on large dataset, or too few Dask workers | Switch to GRNBoost2; create explicit Dask client with more workers |\n| `ModuleNotFoundError: No module named 'arboreto'` | Package not installed | `pip install arboreto` |\n| All importance scores are 0 or identical | Expression matrix has no variance (e.g., all zeros after filtering) | Check data quality; ensure matrix contains non-zero expression values with variance |\n| `TypeError` with NumPy array input | Missing `gene_names` parameter | Provide `gene_names=` list matching the column count of the array |\n\n## Bundled Resources\n\nThis entry is fully self-contained with no `references\u002F` directory.\n\n**Original reference file dispositions**:\n\n- **references\u002Falgorithms.md** (139 lines) -- Algorithm comparison, parameter details, custom regressor kwargs. Consolidated into Key Concepts (Algorithm Selection Guide table) and Key Parameters. Custom regressor kwargs (`regressor_type`, `regressor_kwargs`) omitted: advanced tuning rarely needed; users can pass scikit-learn kwargs directly per the arboreto API docs.\n- **references\u002Fbasic_inference.md** (152 lines) -- Input formats, TF loading, basic workflow, output format. Fully consolidated into Workflow Steps 1-2, Key Concepts (Output Format table), and Quick Start.\n- **references\u002Fdistributed_computing.md** (242 lines) -- Local client, cluster setup, monitoring, performance tips. Consolidated into Workflow Step 3 and Recipe: Distributed Computing. Omitted: Dask dashboard panel descriptions (monitoring UI detail beyond scope), detailed network\u002Fstorage tuning (cluster-admin concern).\n- **scripts\u002Fbasic_grn_inference.py** (98 lines) -- CLI wrapper with argparse. Core load-infer-save logic absorbed into Workflow Steps 1-6 and Quick Start. Argparse\u002FCLI boilerplate stripped.\n\n**Narrative use-case dispositions** (from original Common Use Cases):\n\n- Single-cell RNA-seq analysis -- absorbed into When to Use bullets and Workflow Steps 1\u002F4\n- Bulk RNA-seq with TF filtering -- absorbed into Workflow Step 2 and Quick Start\n- Comparative analysis (multiple conditions) -- omitted as standalone recipe: trivial loop over conditions, pattern shown in Consensus Network recipe\n\n## Related Skills\n\n- **scanpy-scrna-seq** -- upstream single-cell preprocessing (QC, normalization, HVG selection) before GRN inference\n- **scvi-tools-single-cell** -- deep generative models for batch correction and imputation prior to network inference\n- **pyscenic (planned)** -- downstream regulon identification and cellular activity scoring using arboreto adjacencies\n- **networkx-graph-analysis** -- graph-theoretic analysis of inferred regulatory networks (centrality, communities)\n\n## References\n\n- [Arboreto GitHub repository](https:\u002F\u002Fgithub.com\u002Faertslab\u002Farboreto) -- source code, README, examples\n- [Arboreto PyPI](https:\u002F\u002Fpypi.org\u002Fproject\u002Farboreto\u002F) -- installation and version history\n- Moerman et al. (2019) \"GRNBoost2 and Arboreto: efficient and scalable inference of gene regulatory networks\" *Bioinformatics* 35(12):2159-2161 -- algorithm paper\n- Aibar et al. (2017) \"SCENIC: single-cell regulatory network inference and clustering\" *Nature Methods* 14:1083-1086 -- SCENIC pipeline paper\n- [Dask distributed documentation](https:\u002F\u002Fdistributed.dask.org\u002F) -- Dask client, cluster setup, dashboard\n","# Arboreto GRN Inference\n\n## Overview\n\nArboreto infers gene regulatory networks (GRNs) from gene expression data using parallelized tree-based regression. For each target gene, it trains a regression model with all other genes (or a specified TF list) as features and emits TF-target-importance triplets. It provides two interchangeable algorithms -- GRNBoost2 (gradient boosting, fast) and GENIE3 (Random Forest, classic) -- sharing identical input\u002Foutput formats. Computation is Dask-parallelized, scaling from laptop cores to HPC clusters.\n\n## When to Use\n\n- Inferring transcription factor-to-target gene regulatory relationships from bulk RNA-seq expression data\n- Building gene regulatory networks from single-cell RNA-seq count matrices (cells as rows, genes as columns)\n- Generating the adjacency matrix (Step 1) of the pySCENIC regulatory analysis pipeline\n- Comparing regulatory network structure across experimental conditions (e.g., control vs treatment)\n- Producing consensus regulatory networks by running inference across multiple random seeds\n- Validating GRN results by comparing GRNBoost2 and GENIE3 outputs on the same dataset\n- For downstream regulon identification and activity scoring, use arboreto output with pySCENIC\n- For single-cell preprocessing (QC, normalization, clustering) before GRN inference, use **scanpy-scrna-seq**\n\n## Prerequisites\n\n- **Python packages**: `arboreto`, `pandas`, `numpy`, `dask`, `distributed`, `scikit-learn`, `scipy`\n- **Data requirements**: Gene expression matrix (genes as columns, observations as rows) in TSV\u002FCSV; optionally a TF list file (one gene name per line)\n- **Environment**: Python 3.8+; optional `networkx`, `matplotlib` for visualization\n\n```bash\npip install arboreto distributed networkx matplotlib\n```\n\n## Pre-flight Interview\n\nSettle these with the user before writing any analysis code.\n\n```yaml\ndecisions:\n  - id: D1\n    param: algorithm\n    kind: required\n    source: user\n    ask: \"Infer the network with gradient boosting (faster) or random forests (the original GENIE3 formulation)?\"\n    default: \"GRNBoost2\"\n\n  - id: D2\n    param: expressionMatrix\n    kind: required\n    source: upstream\n    ask: \"Which processed expression matrix, and which cells or samples within it, should the network be inferred from?\"\n    default: null\n\n  - id: D3\n    param: regulatorList\n    kind: required\n    source: literature\n    ask: \"Restrict candidate regulators to a curated transcription-factor list, or let every gene be a candidate regulator?\"\n    default: \"a species-matched TF list\"\n\n  - id: D4\n    param: edgeThreshold\n    kind: required\n    source: user\n    ask: \"The output ranks every regulator-target pair - how many of the top edges should be kept as the network?\"\n    default: null\n\n  - id: D5\n    param: seed\n    kind: never_ask\n    source: data\n    reason: \"Fixes the draw for reproducibility; does not change what the data supports\"\n    default: 0\n\n  - id: D6\n    param: daskClient\n    kind: never_ask\n    source: data\n    reason: \"Execution backend; affects runtime only\"\n    default: \"local\"\n```\n\nD3 is the decision that separates a usable network from a hairball. Leaving\nregulators unrestricted lets any correlated gene appear as a regulator, and the\nresult still returns a fully populated ranked table. D4 has no safe default:\nthe algorithm scores all pairs, so where the list is cut is the user's call\nabout the network they intend to interpret.\n\n## Quick Start\n\nComplete GRN inference in a single block. The `if __name__ == '__main__':` guard is required because Dask spawns worker processes via multiprocessing.\n\n```python\nimport pandas as pd\nfrom arboreto.algo import grnboost2\nfrom arboreto.utils import load_tf_names\n\nif __name__ == '__main__':\n    # Load expression matrix (observations x genes)\n    expression_matrix = pd.read_csv('expression_data.tsv', sep='\\t')\n    tf_names = load_tf_names('tf_list.txt')  # optional TF filter\n\n    # Infer GRN (uses all local cores by default)\n    network = grnboost2(expression_data=expression_matrix,\n                        tf_names=tf_names, seed=777)\n\n    # Filter top links and save\n    top_network = network[network['importance'] > 1.0]\n    top_network.to_csv('grn_output.tsv', sep='\\t', index=False, header=False)\n    print(f\"Inferred {len(network)} links, kept {len(top_network)} above threshold\")\n    # Example: Inferred 185432 links, kept 12876 above threshold\n```\n\n## Workflow\n\n### Step 1: Load Expression Data\n\nArboreto accepts a pandas DataFrame (recommended) or NumPy array. Rows are observations (cells or samples), columns are genes. Gene names must be column headers.\n\n```python\nimport pandas as pd\n\n# From TSV (genes as columns, observations as rows)\nexpression_matrix = pd.read_csv('expression_data.tsv', sep='\\t')\nprint(f\"Shape: {expression_matrix.shape}  \"\n      f\"({expression_matrix.shape[0]} observations x {expression_matrix.shape[1]} genes)\")\n# Example: Shape: (5000, 18654)  (5000 observations x 18654 genes)\n\n# From AnnData (e.g., after scanpy preprocessing)\nimport anndata as ad\nadata = ad.read_h5ad('preprocessed.h5ad')\nexpression_matrix = pd.DataFrame(\n    adata.X.toarray() if hasattr(adata.X, 'toarray') else adata.X,\n    columns=adata.var_names.tolist()\n)\nprint(f\"Converted AnnData: {expression_matrix.shape}\")\n# Example: Converted AnnData: (5000, 18654)\n```\n\n### Step 2: Load Transcription Factor List\n\nProviding a TF list restricts regulators to known transcription factors, reducing computation time and improving biological relevance. If omitted, all genes are treated as potential regulators.\n\n```python\nfrom arboreto.utils import load_tf_names\n\n# From file (one TF name per line)\ntf_names = load_tf_names('human_tfs.txt')\nprint(f\"Loaded {len(tf_names)} transcription factors\")\n# Example: Loaded 1639 transcription factors\n\n# Or define directly\ntf_names = ['MYC', 'TP53', 'SOX2', 'NANOG', 'POU5F1']\n\n# Verify TFs exist in expression matrix columns\ntf_in_data = [tf for tf in tf_names if tf in expression_matrix.columns]\nprint(f\"TFs found in expression data: {len(tf_in_data)}\u002F{len(tf_names)}\")\n# Example: TFs found in expression data: 1583\u002F1639\n```\n\n### Step 3: Configure Dask Client (Optional)\n\nBy default arboreto creates an internal Dask client using all local cores. Create an explicit client for resource control, monitoring, or cluster deployment.\n\n```python\nfrom distributed import LocalCluster, Client\n\n# Custom local client with resource limits\nlocal_cluster = LocalCluster(\n    n_workers=8,\n    threads_per_worker=1,   # avoid GIL contention in scikit-learn\n    memory_limit='4GB'       # per worker\n)\nclient = Client(local_cluster)\nprint(f\"Dashboard: {client.dashboard_link}\")\n# Example: Dashboard: http:\u002F\u002F127.0.0.1:8787\u002Fstatus\n```\n\n### Step 4: Run GRN Inference\n\nCall `grnboost2()` (recommended) or `genie3()`. Both share the same signature and output format.\n\n```python\nfrom arboreto.algo import grnboost2\n\nif __name__ == '__main__':\n    network = grnboost2(\n        expression_data=expression_matrix,\n        tf_names=tf_names,         # 'all' if no TF list\n        client_or_address=client,  # omit to use default local scheduler\n        seed=777,                  # for reproducibility\n        verbose=True               # print progress\n    )\n    print(f\"Inferred {len(network)} regulatory links\")\n    print(network.head())\n    # Example:\n    #       TF  target  importance\n    # 0   MYC   CDK4       3.214\n    # 1   MYC   CCND1      2.871\n    # 2  TP53  CDKN1A      2.654\n```\n\n### Step 5: Filter Results\n\nRaw output contains links for every TF-target pair with non-zero importance. Filter to retain high-confidence regulatory interactions.\n\n```python\n# Strategy 1: Importance threshold\nthreshold = 1.0\nfiltered = network[network['importance'] > threshold]\nprint(f\"Threshold {threshold}: {len(filtered)} links \"\n      f\"({len(filtered)\u002Flen(network)*100:.1f}% retained)\")\n# Example: Threshold 1.0: 12876 links (6.9% retained)\n\n# Strategy 2: Top N links per target gene\ntop_n = 10\ntop_per_target = (network.groupby('target')\n                  .apply(lambda g: g.nlargest(top_n, 'importance'))\n                  .reset_index(drop=True))\nprint(f\"Top {top_n} per target: {len(top_per_target)} links\")\n# Example: Top 10 per target: 186540 links\n```\n\n### Step 6: Save Network\n\nSave as a TSV file with three columns: TF, target, importance.\n\n```python\n# Save full network\nnetwork.to_csv('full_network.tsv', sep='\\t', index=False)\nprint(f\"Saved full network: {len(network)} links\")\n\n# Save filtered network (without header for pySCENIC compatibility)\nfiltered.to_csv('filtered_network.tsv', sep='\\t', index=False, header=False)\nprint(f\"Saved filtered network: {len(filtered)} links\")\n\n# Clean up Dask client if explicitly created\nclient.close()\nlocal_cluster.close()\n```\n\n### Step 7: Visualize Results (Optional)\n\nPlot the top regulatory interactions as a directed network graph.\n\n```python\nimport networkx as nx\nimport matplotlib.pyplot as plt\n\n# Build directed graph from top links\ntop_links = network.nlargest(50, 'importance')\nG = nx.from_pandas_edgelist(\n    top_links, source='TF', target='target',\n    edge_attr='importance', create_using=nx.DiGraph()\n)\n\n# Draw network\nfig, ax = plt.subplots(figsize=(12, 10))\npos = nx.spring_layout(G, k=2, seed=42)\nnx.draw_networkx_nodes(G, pos, node_size=300, node_color='lightblue', ax=ax)\nnx.draw_networkx_labels(G, pos, font_size=7, ax=ax)\nedges = nx.draw_networkx_edges(\n    G, pos, edge_color=[G[u][v]['importance'] for u, v in G.edges()],\n    edge_cmap=plt.cm.Reds, width=1.5, arrows=True, ax=ax\n)\nplt.colorbar(edges, ax=ax, label='Importance')\nax.set_title('Top 50 Regulatory Interactions')\nplt.tight_layout()\nplt.savefig('grn_network.png', dpi=300, bbox_inches='tight')\nprint(\"Saved grn_network.png\")\n```\n\n## Key Parameters\n\n| Parameter | Default | Range \u002F Options | Effect |\n|-----------|---------|-----------------|--------|\n| `expression_data` | (required) | DataFrame or ndarray | Expression matrix, observations x genes |\n| `tf_names` | `'all'` | list of str or `'all'` | Restrict regulators to known TFs; `'all'` uses every gene |\n| `gene_names` | `None` | list of str | Required when `expression_data` is a NumPy array |\n| `client_or_address` | `'local'` | Client, str, or `'local'` | Dask client instance or scheduler address |\n| `seed` | `None` | int | Random seed for reproducibility; always set for replicable results |\n| `verbose` | `False` | bool | Print progress messages during inference |\n\n## Key Concepts\n\n### Algorithm Selection Guide\n\nBoth algorithms follow the same multiple-regression strategy (train one model per target gene, extract feature importances) but differ in the underlying regressor.\n\n| Feature | GRNBoost2 | GENIE3 |\n|---------|-----------|--------|\n| **Method** | Stochastic gradient boosting with early stopping | Random Forest (or ExtraTrees) |\n| **Speed** | Fast -- optimized for large datasets | Slower -- higher per-model cost |\n| **Memory** | Lower (early stopping limits tree depth) | Higher (full forests per target) |\n| **Best for** | Default choice; 10k+ observations, single-cell scale | Comparison with published GENIE3 results; validation |\n| **Import** | `from arboreto.algo import grnboost2` | `from arboreto.algo import genie3` |\n| **Output format** | TF, target, importance (identical) | TF, target, importance (identical) |\n\n**Decision rule**: Start with GRNBoost2. Use GENIE3 only when reproducing published GENIE3 analyses or as an independent validation of GRNBoost2 results.\n\n### Output Format\n\nThe network DataFrame has three columns, sorted by descending importance:\n\n| Column | Type | Description |\n|--------|------|-------------|\n| `TF` | str | Transcription factor (regulator) gene name |\n| `target` | str | Target gene name |\n| `importance` | float | Regulatory importance score (higher = stronger predicted regulation) |\n\nImportance scores are not p-values. They represent feature importance from the tree-based model. Filter by absolute threshold or top-N per target for downstream analysis.\n\n### The `if __name__ == '__main__':` Requirement\n\nDask uses Python's `multiprocessing` module to spawn worker processes. Without the main guard, each spawned process re-executes the script, leading to infinite process spawning. Always wrap arboreto calls in `if __name__ == '__main__':` when running as a script. This is not needed inside Jupyter notebooks.\n\n## Common Recipes\n\n### Recipe: Distributed Computing on a Cluster\n\nWhen to use: datasets too large for a single machine (e.g., 50k+ cells, 20k+ genes). Requires a Dask distributed scheduler running on the cluster.\n\n```bash\n# On head node: start scheduler\ndask-scheduler\n# Output: Scheduler at tcp:\u002F\u002F10.118.224.134:8786\n\n# On each compute node: start workers\ndask-worker tcp:\u002F\u002F10.118.224.134:8786 --nprocs 4 --nthreads 1 --memory-limit 16GB\n```\n\n```python\nfrom distributed import Client\nfrom arboreto.algo import grnboost2\nimport pandas as pd\n\nif __name__ == '__main__':\n    client = Client('tcp:\u002F\u002F10.118.224.134:8786')\n    print(f\"Connected to cluster: {client.scheduler_info()['workers'].__len__()} workers\")\n    # Example: Connected to cluster: 16 workers\n\n    expression_data = pd.read_csv('large_scrna.tsv', sep='\\t')\n    tf_names = pd.read_csv('tf_list.txt', header=None)[0].tolist()\n\n    network = grnboost2(\n        expression_data=expression_data,\n        tf_names=tf_names,\n        client_or_address=client,\n        seed=42, verbose=True\n    )\n    network.to_csv('cluster_grn.tsv', sep='\\t', index=False)\n    print(f\"Cluster inference complete: {len(network)} links\")\n    client.close()\n```\n\n### Recipe: Consensus Network from Multiple Seeds\n\nWhen to use: improve robustness by aggregating networks from multiple random seeds. Links appearing consistently across runs are more reliable.\n\n```python\nimport pandas as pd\nfrom distributed import LocalCluster, Client\nfrom arboreto.algo import grnboost2\n\nif __name__ == '__main__':\n    client = Client(LocalCluster(n_workers=8, threads_per_worker=1))\n    expression_data = pd.read_csv('expression_data.tsv', sep='\\t')\n    tf_names = pd.read_csv('tf_list.txt', header=None)[0].tolist()\n\n    seeds = [42, 123, 456, 789, 1001]\n    all_networks = []\n\n    for seed in seeds:\n        net = grnboost2(expression_data=expression_data,\n                        tf_names=tf_names,\n                        client_or_address=client, seed=seed)\n        net['seed'] = seed\n        all_networks.append(net)\n        print(f\"Seed {seed}: {len(net)} links\")\n\n    combined = pd.concat(all_networks, ignore_index=True)\n\n    # Consensus: mean importance across seeds, keep links found in 3+ runs\n    consensus = (combined.groupby(['TF', 'target'])\n                 .agg(mean_importance=('importance', 'mean'),\n                      n_seeds=('seed', 'nunique'))\n                 .reset_index())\n    consensus = consensus[consensus['n_seeds'] >= 3]\n    consensus = consensus.sort_values('mean_importance', ascending=False)\n    consensus.to_csv('consensus_network.tsv', sep='\\t', index=False)\n    print(f\"Consensus network: {len(consensus)} links (found in 3+ of {len(seeds)} runs)\")\n    # Example: Consensus network: 28453 links (found in 3+ of 5 runs)\n\n    client.close()\n```\n\n### Recipe: pySCENIC Integration\n\nWhen to use: downstream regulatory analysis after arboreto GRN inference. Arboreto produces the adjacency matrix (Step 1 of SCENIC); pySCENIC performs regulon identification (Step 2) and activity scoring (Step 3).\n\n```python\nfrom arboreto.algo import grnboost2\nfrom arboreto.utils import load_tf_names\nimport pandas as pd\n\nif __name__ == '__main__':\n    # Step 1 (arboreto): GRN inference\n    expression_matrix = pd.read_csv('scrna_expression.tsv', sep='\\t')\n    tf_names = load_tf_names('allTFs_hg38.txt')\n\n    adjacencies = grnboost2(\n        expression_data=expression_matrix,\n        tf_names=tf_names, seed=777\n    )\n    adjacencies.to_csv('adjacencies.tsv', sep='\\t', index=False, header=False)\n    print(f\"SCENIC Step 1 complete: {len(adjacencies)} adjacencies\")\n    # Example: SCENIC Step 1 complete: 234567 adjacencies\n\n    # Steps 2-3 use pySCENIC CLI (requires pyscenic installation):\n    # pyscenic ctx adjacencies.tsv hg38_ranking.feather \\\n    #     --annotations_fname motifs-v10nr.hgnc-m0.001-o0.0.tbl \\\n    #     --output regulons.csv\n    #\n    # pyscenic aucell scrna_expression.tsv regulons.csv \\\n    #     --output auc_matrix.csv\n    print(\"Next: run pyscenic ctx (Step 2) and pyscenic aucell (Step 3)\")\n```\n\n## Expected Outputs\n\n- `full_network.tsv` -- complete TF-target-importance table; columns: TF, target, importance\n- `filtered_network.tsv` -- high-confidence subset after threshold or top-N filtering\n- `consensus_network.tsv` -- aggregated network from multi-seed runs; columns: TF, target, mean_importance, n_seeds\n- `adjacencies.tsv` -- arboreto output formatted for pySCENIC input (no header)\n- `grn_network.png` -- directed graph visualization of top regulatory interactions\n\n## Troubleshooting\n\n| Problem | Cause | Solution |\n|---------|-------|----------|\n| `RuntimeError: freeze_support()` or infinite process spawning | Missing `if __name__ == '__main__':` guard | Wrap all arboreto calls in main guard; not needed in Jupyter |\n| `MemoryError` during inference | Expression matrix too large for available RAM | Filter low-variance genes first, reduce to top 5-10k genes, or use distributed cluster |\n| Empty or near-empty network | TF names do not match expression matrix column names | Verify TF names overlap with `expression_matrix.columns`; check case sensitivity |\n| Very slow inference | Using GENIE3 on large dataset, or too few Dask workers | Switch to GRNBoost2; create explicit Dask client with more workers |\n| `ModuleNotFoundError: No module named 'arboreto'` | Package not installed | `pip install arboreto` |\n| All importance scores are 0 or identical | Expression matrix has no variance (e.g., all zeros after filtering) | Check data quality; ensure matrix contains non-zero expression values with variance |\n| `TypeError` with NumPy array input | Missing `gene_names` parameter | Provide `gene_names=` list matching the column count of the array |\n\n## Bundled Resources\n\nThis entry is fully self-contained with no `references\u002F` directory.\n\n**Original reference file dispositions**:\n\n- **references\u002Falgorithms.md** (139 lines) -- Algorithm comparison, parameter details, custom regressor kwargs. Consolidated into Key Concepts (Algorithm Selection Guide table) and Key Parameters. Custom regressor kwargs (`regressor_type`, `regressor_kwargs`) omitted: advanced tuning rarely needed; users can pass scikit-learn kwargs directly per the arboreto API docs.\n- **references\u002Fbasic_inference.md** (152 lines) -- Input formats, TF loading, basic workflow, output format. Fully consolidated into Workflow Steps 1-2, Key Concepts (Output Format table), and Quick Start.\n- **references\u002Fdistributed_computing.md** (242 lines) -- Local client, cluster setup, monitoring, performance tips. Consolidated into Workflow Step 3 and Recipe: Distributed Computing. Omitted: Dask dashboard panel descriptions (monitoring UI detail beyond scope), detailed network\u002Fstorage tuning (cluster-admin concern).\n- **scripts\u002Fbasic_grn_inference.py** (98 lines) -- CLI wrapper with argparse. Core load-infer-save logic absorbed into Workflow Steps 1-6 and Quick Start. Argparse\u002FCLI boilerplate stripped.\n\n**Narrative use-case dispositions** (from original Common Use Cases):\n\n- Single-cell RNA-seq analysis -- absorbed into When to Use bullets and Workflow Steps 1\u002F4\n- Bulk RNA-seq with TF filtering -- absorbed into Workflow Step 2 and Quick Start\n- Comparative analysis (multiple conditions) -- omitted as standalone recipe: trivial loop over conditions, pattern shown in Consensus Network recipe\n\n## Related Skills\n\n- **scanpy-scrna-seq** -- upstream single-cell preprocessing (QC, normalization, HVG selection) before GRN inference\n- **scvi-tools-single-cell** -- deep generative models for batch correction and imputation prior to network inference\n- **pyscenic (planned)** -- downstream regulon identification and cellular activity scoring using arboreto adjacencies\n- **networkx-graph-analysis** -- graph-theoretic analysis of inferred regulatory networks (centrality, communities)\n\n## References\n\n- [Arboreto GitHub repository](https:\u002F\u002Fgithub.com\u002Faertslab\u002Farboreto) -- source code, README, examples\n- [Arboreto PyPI](https:\u002F\u002Fpypi.org\u002Fproject\u002Farboreto\u002F) -- installation and version history\n- Moerman et al. (2019) \"GRNBoost2 and Arboreto: efficient and scalable inference of gene regulatory networks\" *Bioinformatics* 35(12):2159-2161 -- algorithm paper\n- Aibar et al. (2017) \"SCENIC: single-cell regulatory network inference and clustering\" *Nature Methods* 14:1083-1086 -- SCENIC pipeline paper\n- [Dask distributed documentation](https:\u002F\u002Fdistributed.dask.org\u002F) -- Dask client, cluster setup, dashboard\n",{"name":5,"license":15,"description":7},"master","fd76f882a04e445e899d5cfb2732f2105ea054ee","skills\u002Fawesome-skills\u002Farboreto-grn-inference",0,"https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS","https:\u002F\u002Fbg-szy.github.io\u002FTOP-SKILLS\u002F","https:\u002F\u002Fgithub.com\u002Fbg-szy","https:\u002F\u002Fwww.ecust.edu.cn\u002F","a student of East China University of Science and Technology","2026-11-15 03:38:02",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add bg-szy\u002FTOP-SKILLS --skill arboreto-grn-inference",{"tool":40,"command":41},"bunx","bunx skills add bg-szy\u002FTOP-SKILLS --skill arboreto-grn-inference",{"tool":43,"command":44},"pnpm","pnpm dlx skills add bg-szy\u002FTOP-SKILLS --skill arboreto-grn-inference","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS\u002Fblob\u002Fmaster\u002Fskills\u002Fawesome-skills\u002Farboreto-grn-inference\u002FSKILL.md",true,[],[49,58,65,72,79,86],{"skillNo":50,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":53,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":57},"01M7DXW9N826PKXK174PVW3038","serpmaven","Scrape Google search results (SERP) to JSON with SERP Maven (serpmaven) - a real Chrome browser on a residential proxy with sticky sessions. Run it as a CLI, a Python library, a local HTTP search API (`serve`), an MCP server or a Docker image; set the proxy login, workers, pages and cost tracking. Use when the user wants Google results, rank tracking, a self-hosted SERP API, or live Google search for an AI agent.","MIT","https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS\u002Fblob\u002Fmaster\u002Fskills\u002Fmarketplace\u002Fserpmaven\u002FSKILL.md",[],[],"2026-11-15 14:19:17",{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":53,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"publishedAt":57},"01M7DXVHSTWDV65475QEQ915M7","revnu-landing-page-audit","Find out why a landing page isn't turning visitors into signups, demos or sales. Audits one page against what the ad, post or search that sends people there promised, then checks the first screen, proof, call to action, friction and mobile. Returns at most 7 prioritized fixes, plus what is fine and what could not be checked; rewrites specific lines and fixes the page in the repo when asked. Never promises a conversion lift. Works without a Revnu account. Use when the user says or asks: \"audit my landing page\", \"why isn't this page converting\", \"review my homepage\", \"my ads get clicks but no signups\", \"people bounce from the pricing page\", \"CRO review\", \"improve conversion rate\", \"is my hero section clear\", \"roast my landing page\", \"rewrite my headline\", \"what's wrong with this page\", \"landing page feedback\", \"make the signup page convert better\". Not for SEO or AI-search audits, page speed engineering on its own, or building a new page from scratch.","https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS\u002Fblob\u002Fmaster\u002Fskills\u002Fmarketplace\u002Frevnu-landing-page-audit\u002FSKILL.md",[],[],{"skillNo":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":53,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":57},"01M7DXV8X2D7B3VTD4HHYYHTGR","revnu-growth-context","Write down once who buys, what they're stuck on, what it costs, what proof you have and what you've already tried, so every growth task starts from facts instead of a questionnaire. Creates or updates .agents\u002Frevnu-growth.md from the repo, the live site and the user, and imports an existing product-marketing.md. Use when the user says \"set up our marketing context\", \"who is our ideal customer\", \"write our ICP\", \"describe the product for marketing\", \"positioning doc\", \"what's our brand voice\", \"remember this about our customers\", \"log what we tried\", \"what have we learned from marketing\", \"we changed pricing, update the context\", \"update our growth notes\", \"summarize the business for the agent\", \"import my product-marketing.md\", \"start a growth doc\", or before a first marketing, SEO, outreach or launch task when no context file exists. Works without a Revnu account.","https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS\u002Fblob\u002Fmaster\u002Fskills\u002Fmarketplace\u002Frevnu-growth-context\u002FSKILL.md",[],[],{"skillNo":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":53,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"publishedAt":57},"01M7DXV0DGJ1G0A60PAHC6J3C5","revnu-get-listed","Get onto the third-party pages AI answers and search cite when buyers ask for the best tool in your category: \"best X\" roundups, directories, review sites, GitHub awesome lists, community threads and comparison pages. Finds which ones are cited, ranks them by fit and effort, and drafts each submission, PR or pitch by the target's own rules. Use when the user says \"get us listed\", \"which directories should we submit to\", \"how do we get into best-of lists\", \"why does ChatGPT recommend competitors and not us\", \"get mentioned in AI answers\", \"submit to awesome lists\", \"get on G2\", \"where are our competitors listed\", \"pitch listicle authors\", \"find backlink opportunities\", \"launch directories\", \"get into roundups\", \"Product Hunt alternatives\", \"where should we be listed\", \"get cited by Perplexity\". Never submits, pays or posts without a yes. Works without a Revnu account.","https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS\u002Fblob\u002Fmaster\u002Fskills\u002Fmarketplace\u002Frevnu-get-listed\u002FSKILL.md",[],[],{"skillNo":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":53,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"publishedAt":57},"01M7DXTSPMGT9D6DSS0HP9BSTA","revnu-ai-readiness-audit","Audit whether your website gives ChatGPT, Perplexity, Claude, Gemini and Google's AI Overviews something worth quoting, and fix the pages that don't. Use when the user says or asks \"is my site ready for AI search\", \"audit my site for AI answers\", \"AEO audit\", \"GEO audit\", \"why doesn't ChatGPT quote our pricing\", \"make our pages citable by AI\", \"do I need llms.txt\", \"does schema help with ChatGPT\", \"check our structured data\", \"is our content visible without JavaScript\", \"AI SEO audit\", \"optimize our homepage for AI answers\", \"LLM SEO check\", \"why does AI get our product wrong\", \"fix our site for AI Overviews\". Checks fetchability and indexability, answers in the raw HTML, entity facts, schema (including JS-injected markup), headings, dates and llms.txt, then gives at most 10 prioritised fixes and edits the repo when asked. Works without a Revnu account.","https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS\u002Fblob\u002Fmaster\u002Fskills\u002Fmarketplace\u002Frevnu-ai-readiness-audit\u002FSKILL.md",[],[],{"skillNo":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":53,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"publishedAt":57},"01M7DXSZHCVTRS620GC3S97JZV","proxy-troubleshooting","Diagnose residential proxy failures, with NodeMaven's gateway as the reference - 407 with correct credentials, 406, 410, 500, a connection that hangs, ERR_TUNNEL_CONNECTION_FAILED in a browser, Playwright or Puppeteer ignoring proxy credentials, a sticky session that changes IP, geo-targeting that is silently ignored, passwords with special characters, and unexpected traffic from headless Chrome. Use when a proxy connection fails, returns the wrong location, or costs more traffic than expected.","https:\u002F\u002Fgithub.com\u002Fbg-szy\u002FTOP-SKILLS\u002Fblob\u002Fmaster\u002Fskills\u002Fmarketplace\u002Fproxy-troubleshooting\u002FSKILL.md",[],[]]