Machine Learning Skills

Use this skill whenever the user wants an end-to-end workflow for the Philadelphia Neurodevelopmental Cohort (PNC) dataset, including BIDS validation, multimodal processing of sMRI, rs-fMRI, task-fMRI, and dMRI, phenotype extraction, and QC integration. Triggers include: 'PNC', 'Philadelphia Neurodevelopmental Cohort', 'process PNC data', 'PNC fMRI', or any request to run the PNC multimodal pipeline.

Use this skill whenever the user wants an end-to-end workflow for the Parkinson's Progression Markers Initiative (PPMI) dataset, including BIDS validation, multimodal processing of sMRI, rs-fMRI, and dMRI, phenotype extraction, and QC integration. Triggers include: 'PPMI', 'Parkinson', 'Parkinson disease', 'process PPMI data', 'PPMI fMRI', or any request to run the PPMI multimodal pipeline.

Use this skill whenever the user wants to run QSIPrep (BIDS App) for diffusion MRI (DWI) preprocessing with best-practice workflows (topup/eddy, denoising/unringing options, susceptibility/motion correction, coregistration/normalization, QC reports) on BIDS datasets. This skill is the NeuroClaw interface-layer wrapper for QSIPrep: it checks installation (Docker/Singularity/conda), generates an execution plan with exact commands and resource estimates, waits for explicit confirmation, then routes all execution through claw-shell.

Use this skill whenever the user wants an end-to-end workflow for the REST-meta-MDD (Resting-State Meta-Major Depressive Disorder) dataset, including BIDS validation, processing of rs-fMRI, phenotype extraction, and QC integration. Triggers include: 'REST-meta-MDD', 'MDD', 'Major Depressive Disorder', 'depression resting-state', 'process REST-meta-MDD', or any request to run the REST-meta-MDD pipeline.

Use this skill whenever the user wants an end-to-end workflow for the SEED-IV (SJTU Emotion EEG Dataset - 4 emotions) dataset, including EEG validation, preprocessing, feature extraction, and emotion classification. Triggers include: 'SEED-IV', 'SEED4', 'emotion EEG', 'EEG emotion recognition', 'process SEED-IV', or any request to run the SEED-IV pipeline.

Use this skill whenever the user wants an end-to-end workflow for the SEED-VIG (SJTU Emotion EEG Dataset - Vigilance) dataset, including EEG validation, preprocessing, feature extraction, and vigilance/fatigue detection. Triggers include: 'SEED-VIG', 'SEEDVIG', 'vigilance EEG', 'fatigue detection', 'drowsiness EEG', 'process SEED-VIG', or any request to run the SEED-VIG pipeline.

Use this skill whenever the user wants an end-to-end workflow for the Transdiagnostic Connectome Project (TCP) dataset, including BIDS validation, multimodal processing of sMRI, rs-fMRI, and dMRI, phenotype extraction, and QC integration. Triggers include: 'TCP', 'Transdiagnostic Connectome', 'process TCP data', 'TCP fMRI', or any request to run the TCP multimodal pipeline.

Use this skill whenever the user wants an end-to-end workflow for the UCLA CNP (Consortium for Neuropsychiatric Phenomics) dataset, including BIDS validation, multimodal processing of sMRI, task-fMRI, and dMRI, phenotype extraction, and QC integration. Triggers include: 'UCLA CNP', 'Consortium Neuropsychiatric Phenomics', 'process UCLA CNP', or any request to run the UCLA CNP multimodal pipeline.

| Craig C. Mello (2006年诺贝尔生理学或医学奖) 的思维框架与决策视角。 核心镜片：简单模型的力量、RNA作为信息货币、跨学科对话。 调研来源：诺奖官网、学术论文、STAT News、NBC News等一手素材。 触发词：「Mello视角」「RNAi思维」「简单模型思维」「基因沉默」「Mello怎么想」。

| 2025年诺贝尔生理学或医学奖得主Fred Ramsdell的思维框架。 聚焦：免疫耐受机制发现、从单基因突变到疾病治疗的全链条思维、工业界科研的价值。 调研来源：7篇一手论文 + Nobel Prize官方资料。信息量有限（极低调的科学家），心智模型基于有限推断。 触发词：FOXP3、Treg、免疫耐受、自身免疫、IPEX、Fred Ramsdell、诺贝尔医学奖2025。

| Robert G. Edwards (1925-2013) 的思维框架与决策模式。2010年诺贝尔生理学或医学奖得主，体外受精(IVF)之父。 基于12个一手/二手来源的深度调研，提炼4个核心心智模型、7条决策启发式和完整的表达DNA。 用途：作为思维顾问，用Edwards的视角分析问题——特别是在科学创新、伦理争议、长期主义和跨学科协作场景中。 当用户提到「用Edwards的视角」「IVF之父怎么看」「Edwards模式」「Robert Edwards perspective」时使用。

| 诺贝尔医学奖得主屠呦呦（Tu Youyou, 2015）的思维框架蒸馏。 核心镜片：从传统智慧中提取科学灵感的"古今转化"思维；低温突破决策。 警告：信息密度低（公开演讲/访谈极少），心智模型基于有限素材推断，诚实边界篇幅大。 触发词：「屠呦呦视角」「青蒿素思维」「古今转化」「传统中药现代化」。

Functional annotation and taxonomy inference from sequence homology.

Aggregate results, train ML models, and produce reports with validated references.

Use omicverse's pyComBat wrapper to remove batch effects from merged bulk RNA-seq or microarray cohorts, export corrected matrices, and benchmark pre/post correction visualisations.

Turn bulk RNA-seq cohorts into synthetic single-cell datasets using omicverse's Bulk2Single workflow for cell fraction estimation, beta-VAE generation, and quality control comparisons against reference scRNA-seq.

Extend scRNA-seq developmental trajectories with BulkTrajBlend by generating intermediate cells from bulk RNA-seq, training beta-VAE and GNN models, and interpolating missing states.

Deep learning framework (PyTorch Lightning). Organize PyTorch code into LightningModules, configure Trainers for multi-GPU/TPU, implement data pipelines, callbacks, logging (W&B, TensorBoard), distributed training (DDP, FSDP, DeepSpeed), for scalable neural network training.

Walk through omicverse's single-cell preprocessing tutorials to QC PBMC3k data, normalise counts, detect HVGs, and run PCA/embedding pipelines on CPU, CPU–GPU mixed, or GPU stacks.

Map scRNA-seq atlases onto spatial transcriptomics slides using omicverse's Single2Spatial workflow for deep-forest training, spot-level assessment, and marker visualisation.

Guide users through omicverse's spatial transcriptomics tutorials covering preprocessing, deconvolution, and downstream modelling workflows across Visium, Visium HD, Stereo-seq, and Slide-seq datasets.

Retrieves gene expression and omics datasets from ArrayExpress and BioStudies with gene disambiguation, experiment quality assessment, and structured reports. Creates comprehensive dataset profiles with metadata, sample information, and download links. Use when users need expression data, omics datasets, or mention ArrayExpress (E-MTAB, E-GEOD) or BioStudies (S-BSST) accessions.

Analyze metabolomics data including metabolite identification, quantification, pathway analysis, and metabolic flux. Processes LC-MS, GC-MS, NMR data from targeted and untargeted experiments. Performs normalization, statistical analysis, pathway enrichment, metabolite-enzyme integration, and biomarker discovery. Use when analyzing metabolomics datasets, identifying differential metabolites, studying metabolic pathways, integrating with transcriptomics/proteomics, discovering metabolic biomarkers, performing flux balance analysis, or characterizing metabolic phenotypes in disease, drug response, or physiological conditions.

Integrate and analyze multiple omics datasets (transcriptomics, proteomics, epigenomics, genomics, metabolomics) for systems biology and precision medicine. Performs cross-omics correlation, multi-omics clustering (MOFA+, NMF), pathway-level integration, and sample matching. Coordinates ToolUniverse skills for expression data (RNA-seq), epigenomics (methylation, ChIP-seq), variants (SNVs, CNVs), protein interactions, and pathway enrichment. Use when analyzing multi-omics datasets, performing integrative analysis, discovering multi-omics biomarkers, studying disease mechanisms across molecular layers, or conducting systems biology research that requires coordinated analysis of transcriptome, genome, epigenome, proteome, and metabolome data.

Production-ready single-cell and expression matrix analysis using scanpy, anndata, and scipy. Performs scRNA-seq QC, normalization, PCA, UMAP, Leiden/Louvain clustering, differential expression (Wilcoxon, t-test, DESeq2), cell type annotation, per-cell-type statistical analysis, gene-expression correlation, batch correction (Harmony), trajectory inference, and cell-cell communication analysis. NEW: Analyzes ligand-receptor interactions between cell types using OmniPath (CellPhoneDB, CellChatDB), scores communication strength, identifies signaling cascades, and handles multi-subunit receptor complexes. Integrates with ToolUniverse gene annotation tools (HPA, Ensembl, MyGene, UniProt) and enrichment tools (gseapy, PANTHER, STRING). Supports h5ad, 10X, CSV/TSV count matrices, and pre-annotated datasets. Use when analyzing single-cell RNA-seq data, studying cell-cell interactions, performing cell type differential expression, computing gene-expression correlations by cell type, analyzing tumor-immune communication, or answering questions about scRNA-seq datasets.

Computational analysis framework for spatial multi-omics data integration. Given spatially variable genes (SVGs), spatial domain annotations, tissue type, and disease context from spatial transcriptomics/proteomics experiments (10x Visium, MERFISH, DBiTplus, SLIDE-seq, etc.), performs comprehensive biological interpretation including pathway enrichment, cell-cell interaction inference, druggable target identification, immune microenvironment characterization, and multi-modal integration. Produces a detailed markdown report with Spatial Omics Integration Score (0-100), domain-by-domain characterization, and validation recommendations. Uses 70+ ToolUniverse tools across 9 analysis phases. Use when users ask about spatial transcriptomics analysis, spatial omics interpretation, tissue heterogeneity, spatial gene expression patterns, tumor microenvironment mapping, tissue zonation, or cell-cell communication from spatial data.

Analyze spatial transcriptomics data to map gene expression in tissue architecture. Supports 10x Visium, MERFISH, seqFISH, Slide-seq, and imaging-based platforms. Performs spatial clustering, domain identification, cell-cell proximity analysis, spatial gene expression patterns, tissue architecture mapping, and integration with single-cell data. Use when analyzing spatial transcriptomics datasets, studying tissue organization, identifying spatial expression patterns, mapping cell-cell interactions in tissue context, characterizing tumor microenvironment spatial structure, or integrating spatial and single-cell RNA-seq data for comprehensive tissue analysis.

Perform statistical modeling and regression analysis on biomedical datasets. Supports linear regression, logistic regression (binary/ordinal/multinomial), mixed-effects models, Cox proportional hazards survival analysis, Kaplan-Meier estimation, and comprehensive model diagnostics. Extracts odds ratios, hazard ratios, confidence intervals, p-values, and effect sizes. Designed to solve BixBench statistical reasoning questions involving clinical/experimental data. Use when asked to fit regression models, compute odds ratios, perform survival analysis, run statistical tests, or interpret model coefficients from provided data.

| dataset understanding + smart downsampling + train/test splits, algorithmic selection (HVG/DE/RF/scGeneFit/SpaPROS), optimal sub-panel discovery (ARI vs size), biological completion with a stability gate (Completion Rule), consensus scoring and completion (only if there is still room), and benchmarking on test splits (ARI/NMI/Silhouette + UMAP similarity).

| annotation, and trajectory inference. These are high-priority actionable workflows — load them first for common single-cell tasks.

Protein language models (ESM3, ESM C) for sequence generation, structure prediction, inverse folding, and embeddings. Design novel proteins, extract ML features, or fold sequences. Local GPU or EvolutionaryScale Forge API. Use AlphaFold for traditional folding; RDKit for small molecules.

Open-source FAIR biology data framework. Version artifacts (AnnData, DataFrame, Zarr), track lineage, validate via ontologies (Bionty), query datasets. Integrates with Nextflow, Snakemake, W&B, scVI. For scRNA-seq use scanpy; for ontology lookups use bionty.

Molecular featurization hub (100+ featurizers) for ML. SMILES to fingerprints (ECFP, MACCS, MAP4), descriptors (RDKit 2D, Mordred), pretrained embeddings (ChemBERTa, GIN, Graphormer), pharmacophores. Scikit-learn compatible with parallelization/caching. For QSAR, virtual screening, similarity, and molecular DL.

> AI autonomous research agent for LLM training optimization using opencode as the agent. The agent autonomously modifies train.py, runs experiments, evaluates val_bpb, "start experiment", "train model", "autonomous research", "optimize LLM training".

Expert agricultural extension professional with 15+ years in farmer training, technology transfer, and rural development. Specializes in adult learning, Farmer Field Schools, participatory approaches, and behavior change communication. Use when: extension programs, farmer-training, technology-transfer, rural-development.

> Expert-level AgentScope developer skill for building production-ready LLM agents. Transforms AI into an experienced AgentScope architect with deep knowledge of ReAct agents, multi-agent orchestration, memory modules, voice agents, MCP/A2A multi-agent, voice agent, MCP, A2A, memory, fine-tuning.

Expert-level AI Application Engineer with deep knowledge of RAG systems, LangChain, LlamaIndex, vector databases, prompt engineering, LLM API integration, and agent frameworks

Expert AI Compute Platform Engineer with 10+ years building and operating large-scale GPU clusters for AI training

Expert LLM Training Engineer with 6+ years of experience in large-scale model pre-training, fine-tuning, alignment, and efficient inference. Use when building, training, or optimizing large language models. Triggers: "llm training", "pre-training", "fine-tuning", "RLHF", "loss spike", "LoRA", "FSDP". Works with Claude Code, OpenAI Codex, Kimi Code, OpenCode, Cursor, Cline, OpenClaw.

Expert-level LLM Research Scientist with deep knowledge of transformer architectures, RLHF, DPO, Constitutional AI, alignment research, evaluation benchmarks, and scaling laws

Elite NLP Engineer skill with expertise in transformer architectures (BERT, GPT, T5), text processing pipelines, LLM fine-tuning, RAG systems, and production NLP deployment. Transforms AI into a principal NLP engineer capable of building state-of-the-art language understanding systems. Use when: nlp, llm, transformers, bert, gpt, text-processing, rag, fine-tuning.

Expert-level Prompt Engineer skill. Transforms AI into a specialist who designs, evaluates, and optimizes prompts for LLMs, RAG pipelines, and agent workflows. Covers prompt patterns (zero-shot, few-shot, CoT, ReAct, Tree-of-Thought), RAG context injection and chunking strategies, agent tool-calling and multi-agent coordination, LLM-as-judge evaluation pipelines, and prompt injection

Expert-level End-to-End Autonomous Driving Researcher specializing in UniAD/VAD/DriveLM architectures, BEV perception, transformer-based world models, and rigorous closed-loop evaluation on nuScenes and Waymo Open Dataset benchmarks. Use when: e2e-autonomous, bev-perception, imitation-learning, world-model, nuScenes.

Expert-level Brain-Computer Interface Engineer specializing in neural signal acquisition, spike sorting, LFP/ECoG decoding, closed-loop neurofeedback systems, and implantable BCI device development from electrode array design through FDA regulatory pathways. Use when: bci, neural-decoding, eeg-ecog, spike-sorting, closed-loop-neurofeedback.

Master calligraphy instructor with 20+ years in Chinese calligraphy (楷书、行书、草书、隶书、篆书) and brush painting. Trained under renowned calligraphers, with expertise in stroke mechanics, character composition, and the philosophical foundations of the art. Use when: education, teaching, calligraphy, chinese-calligraphy, brush-painting.

Expert-level Civil Service Exam Trainer with deep knowledge of government recruitment systems, competitive exam strategies, interview techniques, and career pathway planning for public sector positions