Anton Zhelonkin
Community@tony-zhelonkin · Chicago, Illinois
I`m a MD dermatologist, Computational Biology Postdoc at the University of Chicago. Exploring the gap between science and medicine
Agent Skills by Anton Zhelonkin
Showing 99 vetted skills indexed across 2 GitHub repositories.
tony-zhelonkinanalysis-code-conventions
Routes analysis code decisions to conventions for stages, helpers, dataflow, and R/Python style.
tony-zhelonkinmofa-framework
Routes multi-omics factor analysis tasks across the four MOFA implementations.
tony-zhelonkinte-gene-featurecounts
Generates gene and TE subfamily count matrices from STAR BAMs using containerized featureCounts.
tony-zhelonkinseurat-bridge-integration
Transfer cell-type labels from RNA references onto ATAC queries via multiome bridge integration.
tony-zhelonkinarchitecture-treemap
Render a self-contained HTML treemap visualizing a codebase's architectural model with metrics and coupling edges.
tony-zhelonkinpyranges-peak-gene-linkage
Correlate scATAC peak accessibility with nearby gene expression in paired multiome data.
tony-zhelonkindelegate-cli
Delegate bounded implementation, review, and research tasks to local codex and agy CLIs.
tony-zhelonkinte-reference-saf-build
Builds shared TE reference SAF files from TEtranscripts RepeatMasker GTFs for RNA-seq counting.
tony-zhelonkinpeak-atlas-unpaired
Builds consensus scATAC peak atlases from unpaired RNA and ATAC data via label transfer.
tony-zhelonkinstar-te-preprocessing
Configures STAR alignment and featureCounts to retain transposable-element signal in bulk RNA-seq counts.
tony-zhelonkinscired
Interprets scRNA-seq factor analysis via Poisson-GLM residualisation, PCA, varimax rotation, and factor-covariate scoring.
tony-zhelonkinannotate-bulk-rnaseq-data
Annotate bulk RNA-seq featureCounts matrices with gene symbols and TE families before edgeR/limma DE.
tony-zhelonkinsignac-chromatin-analysis
Analyzes 10x Multiome scATAC-seq data in R with Signac, from TF-IDF and LSI to ChromVAR motif enrichment.
tony-zhelonkinreasoning-trace
Persist analysis decisions and findings as git-tracked Markdown notes.
tony-zhelonkinmentor-mode
Guides learning-preserving explanations and stepwise bioinformatics coding sessions.
tony-zhelonkinmofa-mofax
Inspect trained MOFA+ HDF5 models in Python with factor, weight, and variance accessors.
tony-zhelonkinnotebook-annotation
Scaffolds a marimo and jscatter notebook for multi-round single-cell relabeling campaigns with selection manifests.
tony-zhelonkinmofa-r
Fit MOFA2 multi-omics factor models from R and plot factors, weights, and variance explained.
tony-zhelonkinmllmcelltype-consensus-annotation
Annotates scRNA-seq clusters via multi-LLM consensus over marker genes with Python-recomputed confidence metrics.
tony-zhelonkinharmonypy-batch-integration
Correct batch effects on PCA, LSI, or spectral embeddings using harmonypy via scanpy.external.harmony_integrate.
tony-zhelonkinsnapatac2-atac-preprocessing
Preprocess scATAC-seq fragment files into embeddings, peak matrices, and gene-activity scores with SnapATAC2.
tony-zhelonkinconsensus-nmf-multirun
Runs multi-variant consensus cNMF to discover robust gene programs in single-cell RNA-seq data.
tony-zhelonkinmuon-multimodal-analysis
Orchestrates paired scRNA and scATAC multimodal analysis workflows using muon, scanpy, and SnapATAC2.
tony-zhelonkinnotebook-exploration
Scaffold read-only exploration notebooks for single-cell analysis results using jscatter or Quarto.
Frequently Asked Questions About Anton Zhelonkin
FAQPage SchemaWhat tasks can I perform with tony-zhelonkin's skill registry?▼
You can run end-to-end single-cell and bulk genomics analyses: scRNA-seq QC, clustering, and batch correction (Scanpy, scVI, Harmony); scATAC-seq preprocessing, peak atlas construction, and motif enrichment (SnapATAC2, Signac, chromVAR); multiome integration (MultiVI, scGLUE, Seurat bridge); TF footprinting (TOBIAS, HINT-ATAC); and bulk RNA-seq GSEA, TF/pathway activity inference, and TE quantification.
Who is the target audience for these skills?▼
Computational biologists, bioinformaticians, and wet-lab-adjacent researchers analyzing single-cell or bulk sequencing data. The author is an MD dermatologist and Computational Biology Postdoc at the University of Chicago, and the skills encode house conventions for reproducible, restartable analysis pipelines in R and Python.
How do the skills handle multi-omics integration?▼
Paired 10x Multiome data routes through muon/MuData, scVI MultiVI, or Seurat+Signac WNN; unpaired RNA/ATAC uses scGLUE or Seurat gene-activity anchoring; ATAC-to-RNA label transfer uses bridge integration. MOFA-family skills (mofapy2, MOFA2, MOFAcellulaR, mofax) cover Bayesian multi-view factor analysis.
Are these skills open source and what do they cost?▼
The skills carry MIT licenses per their native frontmatter and are free to use. They orchestrate open-source packages (scanpy, scvi-tools, Seurat, Signac, nf-core/rnaseq, pyGenomeTracks), so runtime cost is only your compute—some skills note GPU/SVI options and Docker-based execution.
What prerequisites and data inputs do the skills require?▼
Typical inputs are 10x FASTQs or fragments.tsv.gz, CellRanger/CellRanger-ARC outputs, nf-core/rnaseq BAMs, or .h5ad/.rds objects. Skills assume R (Seurat/Signac/Bioconductor) and Python (scverse, scvi-tools) environments, Docker for nf-core and hosting skills, and reference assets like MSigDB, CollecTRI, or TEtranscripts RepeatMasker GTFs.