Gget

gget

Unified CLI and Python access to 20+ genomic databases. All modules work as both command-line tools and Python functions.

Installation

uv pip install --upgrade gget

Some modules require setup: gget setup alphafold|cellxgene|elm|gpt

Quick Start

# CLI: gget <module> [arguments]
gget search -s human BRCA1
gget info ENSG00000012048
gget seq ENSG00000012048 -t   # protein sequence

# Python: gget.module(arguments)
import gget
gget.search(["BRCA1"], species="homo_sapiens")
gget.info(["ENSG00000012048"])

Common flags: -o (save to file), -csv (CSV output), -q (quiet)

Supporting Files

• [module_reference.md](references/module_reference.md) - Complete parameter reference for all 20+ modules
• [database_info.md](references/database_info.md) - Database descriptions and update frequencies
• [workflows.md](references/workflows.md) - Extended workflow examples

Scripts

• `scripts/gene_analysis.py` - Gene discovery to sequence analysis pipeline
• `scripts/enrichment_pipeline.py` - Gene list enrichment workflow
• `scripts/batch_sequence_analysis.py` - Batch BLAST/alignment processing

Module Overview

Reference & Gene Information

Sequence Analysis

Structure & Protein

Expression & Disease

Other

Key Workflows

Gene Discovery → Sequence Analysis

# Search → info → sequence → BLAST
results = gget.search(["GABA", "receptor"], species="homo_sapiens")
info = gget.info(results["ensembl_id"].tolist()[:5])
sequences = gget.seq(results["ensembl_id"].tolist()[:5], translate=True)
blast_hits = gget.blast(my_sequence, database="swissprot", limit=10)

Expression & Enrichment

# Tissue expression → correlated genes → enrichment
tissue_expr = gget.archs4("ACE2", which="tissue")
correlated = gget.archs4("ACE2", which="correlation")
enrichment = gget.enrichr(correlated["gene_symbol"].tolist()[:50], database="ontology", plot=True)

Enrichr Database Shortcuts

Single-Cell Data

# Gene symbols are case-sensitive: 'PAX7' (human), 'Pax7' (mouse)
adata = gget.cellxgene(gene=["ACE2", "ABCA1"], tissue="lung", cell_type="epithelial cell")
# Filters: disease, development_stage, sex, assay, donor_id, ethnicity

Comparative Genomics

orthologs = gget.bgee("ENSG00000169194", type="orthologs")
human_seq = gget.seq("ENSG00000169194", translate=True)
alignment = gget.muscle([human_seq, mouse_seq])

Best Practices

• Use --limit to control result sizes
• Save results with -o for reproducibility
• Process max ~1000 Ensembl IDs at once with gget info
• Use gget diamond with --threads for faster local alignment; save DB with --diamond_db
• For gget muscle, use -s5 (Super5) for large datasets
• AlphaFold multimer: use -mr 20 for accuracy, -r for AMBER relaxation
• Update regularly: uv pip install --upgrade gget (databases change structure)

Attribution

Adapted from K-Dense-AI/claude-scientific-skills (BSD-2-Clause). Citation: Luebbert & Pachter (2023) Bioinformatics. https://doi.org/10.1093/bioinformatics/btac836

gget Database Information

Overview of databases queried by gget modules, including update frequencies and important considerations.

Important Note

The databases queried by gget are continuously being updated, which sometimes changes their structure. gget modules are tested automatically on a biweekly basis and updated to match new database structures when necessary. Always keep gget updated:

pip install --upgrade gget

Database Directory

Genomic Reference Databases

Ensembl

• Used by: gget ref, gget search, gget info, gget seq
• Description: Comprehensive genome database with annotations for vertebrate and invertebrate species
• Update frequency: Regular releases (numbered); new releases approximately every 3 months
• Access: FTP downloads, REST API
• Website: https://www.ensembl.org/
• Notes:
• Supports both vertebrate and invertebrate genomes
• Can specify release number for reproducibility
• Shortcuts available for common species ('human', 'mouse')

UCSC Genome Browser

• Used by: gget blat
• Description: Genome browser database with BLAT alignment tool
• Update frequency: Regular updates with new assemblies
• Access: Web service API
• Website: https://genome.ucsc.edu/
• Notes:
• Multiple genome assemblies available (hg38, mm39, etc.)
• BLAT optimized for vertebrate genomes

Protein & Structure Databases

UniProt

• Used by: gget info, gget seq (amino acid sequences), gget elm
• Description: Universal Protein Resource, comprehensive protein sequence and functional information
• Update frequency: Regular releases (weekly for Swiss-Prot, monthly for TrEMBL)
• Access: REST API
• Website: https://www.uniprot.org/
• Notes:
• Swiss-Prot: manually annotated and reviewed
• TrEMBL: automatically annotated

NCBI (National Center for Biotechnology Information)

• Used by: gget info, gget bgee (for non-Ensembl species)
• Description: Gene and protein databases with extensive cross-references
• Update frequency: Continuous updates
• Access: E-utilities API
• Website: https://www.ncbi.nlm.nih.gov/
• Databases: Gene, Protein, RefSeq

RCSB PDB (Protein Data Bank)

• Used by: gget pdb
• Description: Repository of 3D structural data for proteins and nucleic acids
• Update frequency: Weekly updates
• Access: REST API
• Website: https://www.rcsb.org/
• Notes:
• Experimentally determined structures (X-ray, NMR, cryo-EM)
• Includes metadata about experiments and publications

ELM (Eukaryotic Linear Motif)

• Used by: gget elm
• Description: Database of functional sites in eukaryotic proteins
• Update frequency: Periodic updates
• Access: Downloaded database (via gget setup elm)
• Website: http://elm.eu.org/
• Notes:
• Requires local download before first use
• Contains validated motifs and patterns

Sequence Similarity Databases

BLAST Databases (NCBI)

• Used by: gget blast
• Description: Pre-formatted databases for BLAST searches
• Update frequency: Regular updates
• Access: NCBI BLAST API
• Databases:
• Nucleotide: nt (all GenBank), refseq_rna, pdbnt
• Protein: nr (non-redundant), swissprot, pdbaa, refseq_protein
• Notes:
• nt and nr are very large databases
• Consider specialized databases for faster, more focused searches

Expression & Correlation Databases

ARCHS4

• Used by: gget archs4
• Description: Massive mining of publicly available RNA-seq data
• Update frequency: Periodic updates with new samples
• Access: HTTP API
• Website: https://maayanlab.cloud/archs4/
• Data:
• Human and mouse RNA-seq data
• Correlation matrices
• Tissue expression atlases
• Citation: Lachmann et al., Nature Communications, 2018

CZ CELLxGENE Discover

• Used by: gget cellxgene
• Description: Single-cell RNA-seq data from multiple studies
• Update frequency: Continuous additions of new datasets
• Access: Census API (via cellxgene-census package)
• Website: https://cellxgene.cziscience.com/
• Data:
• Single-cell RNA-seq count matrices
• Cell type annotations
• Tissue and disease metadata
• Notes:
• Requires gget setup cellxgene
• Gene symbols are case-sensitive
• May not support latest Python versions

Bgee

• Used by: gget bgee
• Description: Gene expression and orthology database
• Update frequency: Regular releases
• Access: REST API
• Website: https://www.bgee.org/
• Data:
• Gene expression across tissues and developmental stages
• Orthology relationships across species
• Citation: Bastian et al., 2021

Functional & Pathway Databases

Enrichr / modEnrichr

• Used by: gget enrichr
• Description: Gene set enrichment analysis web service
• Update frequency: Regular updates to underlying databases
• Access: REST API
• Website: https://maayanlab.cloud/Enrichr/
• Databases included:
• KEGG pathways
• Gene Ontology (GO)
• Transcription factor targets (ChEA)
• Disease associations (GWAS Catalog)
• Cell type markers (PanglaoDB)
• Notes:
• Supports multiple model organisms
• Background gene lists can be provided for custom enrichment

Disease & Drug Databases

Open Targets

• Used by: gget opentargets
• Description: Integrative platform for disease-target associations
• Update frequency: Regular releases (quarterly)
• Access: GraphQL API
• Website: https://www.opentargets.org/
• Data:
• Disease associations
• Drug information and clinical trials
• Target tractability
• Pharmacogenetics
• Gene expression
• DepMap gene-disease effects
• Protein-protein interactions

cBioPortal

• Used by: gget cbio
• Description: Cancer genomics data portal
• Update frequency: Continuous addition of new studies
• Access: Web API, downloadable datasets
• Website: https://www.cbioportal.org/
• Data:
• Mutations, copy number alterations, structural variants
• Gene expression
• Clinical data
• Notes:
• Large datasets; caching recommended
• Multiple cancer types and studies available

COSMIC (Catalogue Of Somatic Mutations In Cancer)

• Used by: gget cosmic
• Description: Comprehensive cancer mutation database
• Update frequency: Regular releases
• Access: Download (requires account and license for commercial use)
• Website: https://cancer.sanger.ac.uk/cosmic
• Data:
• Somatic mutations in cancer
• Gene census
• Cell line data
• Drug resistance mutations
• Important:
• Free for academic use
• License fees apply for commercial use
• Requires COSMIC account credentials
• Must download database before querying

AI & Prediction Services

AlphaFold2 (DeepMind)

• Used by: gget alphafold
• Description: Deep learning model for protein structure prediction
• Model version: Simplified version for local execution
• Access: Local computation (requires model download via gget setup)
• Website: https://alphafold.ebi.ac.uk/
• Notes:
• Requires ~4GB model parameters download
• Requires OpenMM installation
• Computationally intensive
• Python version-specific requirements

OpenAI API

• Used by: gget gpt
• Description: Large language model API
• Update frequency: New models released periodically
• Access: REST API (requires API key)
• Website: https://openai.com/
• Notes:
• Default model: gpt-3.5-turbo
• Free tier limited to 3 months after account creation
• Set billing limits to control costs

Data Consistency & Reproducibility

Version Control

To ensure reproducibility in analyses:

1. Specify database versions/releases:

   # Use specific Ensembl release
   gget.ref("homo_sapiens", release=110)

   # Use specific Census version
   gget.cellxgene(gene=["PAX7"], census_version="2023-07-25")

2. Document gget version:

   import gget
   print(gget.__version__)

3. Save raw data:

   # Always save results for reproducibility
   results = gget.search(["ACE2"], species="homo_sapiens")
   results.to_csv("search_results_2025-01-15.csv", index=False)

Handling Database Updates

1. Regular gget updates:

• Update gget biweekly to match database structure changes
• Check release notes for breaking changes

2. Error handling:

• Database structure changes may cause temporary failures
• Check GitHub issues: https://github.com/pachterlab/gget/issues
• Update gget if errors occur

3. API rate limiting:

• Implement delays for large-scale queries
• Use local databases (DIAMOND, COSMIC) when possible
• Cache results to avoid repeated queries

Database-Specific Best Practices

Ensembl

• Use species shortcuts ('human', 'mouse') for convenience
• Specify release numbers for reproducibility
• Check available species with gget ref --list_species

UniProt

• UniProt IDs are more stable than gene names
• Swiss-Prot annotations are manually curated and more reliable
• Use PDB flag in gget info only when needed (increases runtime)

BLAST/BLAT

• Start with default parameters, then optimize
• Use specialized databases (swissprot, refseq_protein) for focused searches
• Consider E-value cutoffs based on query length

Expression Databases

• Gene symbols are case-sensitive in CELLxGENE
• ARCHS4 correlation data is based on co-expression patterns
• Consider tissue-specificity when interpreting results

Cancer Databases

• cBioPortal: cache data locally for repeated analyses
• COSMIC: download appropriate database subset for your needs
• Respect license agreements for commercial use

Citations

When using gget, cite both the gget publication and the underlying databases:

gget: Luebbert, L. & Pachter, L. (2023). Efficient querying of genomic reference databases with gget. Bioinformatics. https://doi.org/10.1093/bioinformatics/btac836

Database-specific citations: Check references/ directory or database websites for appropriate citations.

gget Module Reference

Comprehensive parameter reference for all gget modules.

Reference & Gene Information Modules

gget ref

Retrieve Ensembl reference genome FTPs and metadata.

Parameters:

Returns: JSON containing FTP links, Ensembl release numbers, release dates, file sizes

---

gget search

Search for genes by name or description in Ensembl.

Parameters:

Returns: ensembl_id, gene_name, ensembl_description, ext_ref_description, biotype, URL

---

gget info

Get comprehensive gene/transcript metadata from Ensembl, UniProt, and NCBI.

Parameters:

Python-specific:

• save=True: Save output to current directory
• wrap_text=True: Format dataframe with wrapped text

Note: Processing >1000 IDs simultaneously may cause server errors.

Returns: UniProt ID, NCBI gene ID, gene name, synonyms, protein names, descriptions, biotype, canonical transcript

---

gget seq

Retrieve nucleotide or amino acid sequences in FASTA format.

Parameters:

Data sources: Ensembl (nucleotide), UniProt (amino acid)

Returns: FASTA format sequences

---

Sequence Analysis & Alignment Modules

gget blast

BLAST sequences against standard databases.

Parameters:

Returns: Description, Scientific Name, Common Name, Taxid, Max Score, Total Score, Query Coverage

---

gget blat

Find genomic positions using UCSC BLAT.

Parameters:

Returns: genome, query size, alignment start/end, matches, mismatches, alignment percentage

---

gget muscle

Align multiple sequences using Muscle5.

Parameters:

Returns: ClustalW format alignment or aligned FASTA (.afa)

---

gget diamond

Fast local protein/translated DNA alignment.

Parameters:

Returns: Identity %, sequence lengths, match positions, gap openings, E-values, bit scores

---

Structural & Protein Analysis Modules

gget pdb

Query RCSB Protein Data Bank.

Parameters:

Returns: PDB format (structures) or JSON (metadata)

---

gget alphafold

Predict 3D protein structures using AlphaFold2.

Setup: Requires OpenMM and gget setup alphafold (~4GB download)

Parameters:

Python-only:

• plot (bool): Generate 3D visualization (default: True)
• show_sidechains (bool): Include side chains (default: True)

Note: Multiple sequences automatically trigger multimer modeling

Returns: PDB structure file, JSON alignment error data, optional 3D plot

---

gget elm

Predict Eukaryotic Linear Motifs.

Setup: Requires gget setup elm

Parameters:

Returns: Two outputs: 1. ortholog_df: Motifs from orthologous proteins 2. regex_df: Motifs matched in input sequence

---

Expression & Disease Data Modules

gget archs4

Query ARCHS4 for gene correlation or tissue expression.

Parameters:

Returns:

• correlation: Gene symbols, Pearson correlation coefficients (top 100)
• tissue: Tissue IDs, min/Q1/median/Q3/max expression

---

gget cellxgene

Query CZ CELLxGENE Discover Census for single-cell data.

Setup: Requires gget setup cellxgene

Parameters:

Additional filters: disease, development_stage, sex, assay, dataset_id, donor_id, ethnicity, suspension_type

Important: Gene symbols are case-sensitive ('PAX7' for human, 'Pax7' for mouse)

Returns: AnnData object with count matrices and metadata

---

gget enrichr

Perform enrichment analysis using Enrichr/modEnrichr.

Parameters:

Python-only:

• plot (bool): Generate graphical results

Database shortcuts:

• 'pathway' → KEGG_2021_Human
• 'transcription' → ChEA_2016
• 'ontology' → GO_Biological_Process_2021
• 'diseases_drugs' → GWAS_Catalog_2019
• 'celltypes' → PanglaoDB_Augmented_2021

Returns: Pathway/function associations with adjusted p-values, overlapping gene counts

---

gget bgee

Retrieve orthology and expression from Bgee.

Parameters:

Note: Multiple IDs supported when type='expression'

Returns:

• orthologs: Genes across species with IDs, names, taxonomic info
• expression: Anatomical entities, confidence scores, expression status

---

gget opentargets

Retrieve disease/drug associations from OpenTargets.

Parameters:

Resource-specific filters:

• drugs: --filter_disease
• pharmacogenetics: --filter_drug
• expression/depmap: --filter_tissue, --filter_anat_sys, --filter_organ
• interactions: --filter_protein_a, --filter_protein_b, --filter_gene_b

Returns: Disease/drug associations, tractability, pharmacogenetics, expression, DepMap, interactions

---

gget cbio

Plot cancer genomics heatmaps from cBioPortal.

Subcommands: search, plot

search parameters:

plot parameters:

Returns: PNG heatmap figure

---

gget cosmic

Search COSMIC database for cancer mutations.

Important: License fees for commercial use. Requires COSMIC account.

Query parameters:

Download parameters:

Note: First-time users must download database

Returns: Mutation data from COSMIC

---

Additional Tools

gget mutate

Generate mutated nucleotide sequences.

Parameters:

Returns: Mutated sequences in FASTA format

---

gget gpt

Generate text using OpenAI's API.

Setup: Requires gget setup gpt and OpenAI API key

Parameters:

Important: Free tier limited to 3 months. Set billing limits.

Returns: Generated text string

---

gget setup

Install/download dependencies for modules.

Parameters:

Modules requiring setup:

• alphafold - Downloads ~4GB model parameters
• cellxgene - Installs cellxgene-census
• elm - Downloads local ELM database
• gpt - Configures OpenAI integration

Returns: None (installs dependencies)

gget Workflow Examples

Extended workflow examples demonstrating how to combine multiple gget modules for common bioinformatics tasks.

Table of Contents

1. Complete Gene Analysis Pipeline 2. Comparative Structural Biology 3. Cancer Genomics Analysis 4. Single-Cell Expression Analysis 5. Building Reference Transcriptomes 6. Mutation Impact Assessment 7. Drug Target Discovery

---

Complete Gene Analysis Pipeline

Comprehensive analysis of a gene from discovery to functional annotation.

import gget
import pandas as pd

# Step 1: Search for genes of interest
print("Step 1: Searching for GABA receptor genes...")
search_results = gget.search(["GABA", "receptor", "alpha"],
                             species="homo_sapiens",
                             andor="and")
print(f"Found {len(search_results)} genes")

# Step 2: Get detailed information
print("\nStep 2: Getting detailed information...")
gene_ids = search_results["ensembl_id"].tolist()[:5]  # Top 5 genes
gene_info = gget.info(gene_ids, pdb=True)
print(gene_info[["ensembl_id", "gene_name", "uniprot_id", "description"]])

# Step 3: Retrieve sequences
print("\nStep 3: Retrieving sequences...")
nucleotide_seqs = gget.seq(gene_ids)
protein_seqs = gget.seq(gene_ids, translate=True)

# Save sequences
with open("gaba_receptors_nt.fasta", "w") as f:
    f.write(nucleotide_seqs)
with open("gaba_receptors_aa.fasta", "w") as f:
    f.write(protein_seqs)

# Step 4: Get expression data
print("\nStep 4: Getting tissue expression...")
for gene_id, gene_name in zip(gene_ids, gene_info["gene_name"]):
    expr_data = gget.archs4(gene_name, which="tissue")
    print(f"\n{gene_name} expression:")
    print(expr_data.head())

# Step 5: Find correlated genes
print("\nStep 5: Finding correlated genes...")
correlated = gget.archs4(gene_info["gene_name"].iloc[0], which="correlation")
correlated_top = correlated.head(20)
print(correlated_top)

# Step 6: Enrichment analysis on correlated genes
print("\nStep 6: Performing enrichment analysis...")
gene_list = correlated_top["gene_symbol"].tolist()
enrichment = gget.enrichr(gene_list, database="ontology", plot=True)
print(enrichment.head(10))

# Step 7: Get disease associations
print("\nStep 7: Getting disease associations...")
for gene_id, gene_name in zip(gene_ids[:3], gene_info["gene_name"][:3]):
    diseases = gget.opentargets(gene_id, resource="diseases", limit=5)
    print(f"\n{gene_name} disease associations:")
    print(diseases)

# Step 8: Check for orthologs
print("\nStep 8: Finding orthologs...")
orthologs = gget.bgee(gene_ids[0], type="orthologs")
print(orthologs)

print("\nComplete gene analysis pipeline finished!")

---

Comparative Structural Biology

Compare protein structures across species and analyze functional motifs.

import gget

# Define genes for comparison
human_gene = "ENSG00000169174"  # PCSK9
mouse_gene = "ENSMUSG00000044254"  # Pcsk9

print("Comparative Structural Biology Workflow")
print("=" * 50)

# Step 1: Get gene information
print("\n1. Getting gene information...")
human_info = gget.info([human_gene])
mouse_info = gget.info([mouse_gene])

print(f"Human: {human_info['gene_name'].iloc[0]}")
print(f"Mouse: {mouse_info['gene_name'].iloc[0]}")

# Step 2: Retrieve protein sequences
print("\n2. Retrieving protein sequences...")
human_seq = gget.seq(human_gene, translate=True)
mouse_seq = gget.seq(mouse_gene, translate=True)

# Save to file for alignment
with open("pcsk9_sequences.fasta", "w") as f:
    f.write(human_seq)
    f.write("\n")
    f.write(mouse_seq)

# Step 3: Align sequences
print("\n3. Aligning sequences...")
alignment = gget.muscle("pcsk9_sequences.fasta")
print("Alignment completed. Visualizing in ClustalW format:")
print(alignment)

# Step 4: Get existing structures from PDB
print("\n4. Searching PDB for existing structures...")
# Search by sequence using BLAST
pdb_results = gget.blast(human_seq, database="pdbaa", limit=5)
print("Top PDB matches:")
print(pdb_results[["Description", "Max Score", "Query Coverage"]])

# Download top structure
if len(pdb_results) > 0:
    # Extract PDB ID from description (usually format: "PDB|XXXX|...")
    pdb_id = pdb_results.iloc[0]["Description"].split("|")[1]
    print(f"\nDownloading PDB structure: {pdb_id}")
    gget.pdb(pdb_id, save=True)

# Step 5: Predict AlphaFold structures
print("\n5. Predicting structures with AlphaFold...")
# Note: This requires gget setup alphafold and is computationally intensive
# Uncomment to run:
# human_structure = gget.alphafold(human_seq, plot=True)
# mouse_structure = gget.alphafold(mouse_seq, plot=True)
print("(AlphaFold prediction skipped - uncomment to run)")

# Step 6: Identify functional motifs
print("\n6. Identifying functional motifs with ELM...")
# Note: Requires gget setup elm
# Uncomment to run:
# human_ortholog_df, human_regex_df = gget.elm(human_seq)
# print("Human PCSK9 functional motifs:")
# print(human_regex_df)
print("(ELM analysis skipped - uncomment to run)")

# Step 7: Get orthology information
print("\n7. Getting orthology information from Bgee...")
orthologs = gget.bgee(human_gene, type="orthologs")
print("PCSK9 orthologs:")
print(orthologs)

print("\nComparative structural biology workflow completed!")

---

Cancer Genomics Analysis

Analyze cancer-associated genes and their mutations.

import gget
import matplotlib.pyplot as plt

print("Cancer Genomics Analysis Workflow")
print("=" * 50)

# Step 1: Search for cancer-related genes
print("\n1. Searching for breast cancer genes...")
genes = gget.search(["breast", "cancer", "BRCA"],
                    species="homo_sapiens",
                    andor="or",
                    limit=20)
print(f"Found {len(genes)} genes")

# Focus on specific genes
target_genes = ["BRCA1", "BRCA2", "TP53", "PIK3CA", "ESR1"]
print(f"\nAnalyzing: {', '.join(target_genes)}")

# Step 2: Get gene information
print("\n2. Getting gene information...")
gene_search = []
for gene in target_genes:
    result = gget.search([gene], species="homo_sapiens", limit=1)
    if len(result) > 0:
        gene_search.append(result.iloc[0])

gene_df = pd.DataFrame(gene_search)
gene_ids = gene_df["ensembl_id"].tolist()

# Step 3: Get disease associations
print("\n3. Getting disease associations from OpenTargets...")
for gene_id, gene_name in zip(gene_ids, target_genes):
    print(f"\n{gene_name} disease associations:")
    diseases = gget.opentargets(gene_id, resource="diseases", limit=3)
    print(diseases[["disease_name", "overall_score"]])

# Step 4: Get drug associations
print("\n4. Getting drug associations...")
for gene_id, gene_name in zip(gene_ids[:3], target_genes[:3]):
    print(f"\n{gene_name} drug associations:")
    drugs = gget.opentargets(gene_id, resource="drugs", limit=3)
    if len(drugs) > 0:
        print(drugs[["drug_name", "drug_type", "max_phase_for_all_diseases"]])

# Step 5: Search cBioPortal for studies
print("\n5. Searching cBioPortal for breast cancer studies...")
studies = gget.cbio_search(["breast", "cancer"])
print(f"Found {len(studies)} studies")
print(studies[:5])

# Step 6: Create cancer genomics heatmap
print("\n6. Creating cancer genomics heatmap...")
if len(studies) > 0:
    # Select relevant studies
    selected_studies = studies[:2]  # Top 2 studies

    gget.cbio_plot(
        selected_studies,
        target_genes,
        stratification="cancer_type",
        variation_type="mutation_occurrences",
        show=False
    )
    print("Heatmap saved to ./gget_cbio_figures/")

# Step 7: Query COSMIC database (requires setup)
print("\n7. Querying COSMIC database...")
# Note: Requires COSMIC account and database download
# Uncomment to run:
# for gene in target_genes[:2]:
#     cosmic_results = gget.cosmic(
#         gene,
#         cosmic_tsv_path="cosmic_cancer.tsv",
#         limit=10
#     )
#     print(f"\n{gene} mutations in COSMIC:")
#     print(cosmic_results)
print("(COSMIC query skipped - requires database download)")

# Step 8: Enrichment analysis
print("\n8. Performing pathway enrichment...")
enrichment = gget.enrichr(target_genes, database="pathway", plot=True)
print("\nTop enriched pathways:")
print(enrichment.head(10))

print("\nCancer genomics analysis completed!")

---

Single-Cell Expression Analysis

Analyze single-cell RNA-seq data for specific cell types and tissues.

import gget
import scanpy as sc

print("Single-Cell Expression Analysis Workflow")
print("=" * 50)

# Note: Requires gget setup cellxgene

# Step 1: Define genes and cell types of interest
genes_of_interest = ["ACE2", "TMPRSS2", "CD4", "CD8A"]
tissue = "lung"
cell_types = ["type ii pneumocyte", "macrophage", "t cell"]

print(f"\nAnalyzing genes: {', '.join(genes_of_interest)}")
print(f"Tissue: {tissue}")
print(f"Cell types: {', '.join(cell_types)}")

# Step 2: Get metadata first
print("\n1. Retrieving metadata...")
metadata = gget.cellxgene(
    gene=genes_of_interest,
    tissue=tissue,
    species="homo_sapiens",
    meta_only=True
)
print(f"Found {len(metadata)} datasets")
print(metadata.head())

# Step 3: Download count matrices
print("\n2. Downloading single-cell data...")
# Note: This can be a large download
adata = gget.cellxgene(
    gene=genes_of_interest,
    tissue=tissue,
    species="homo_sapiens",
    census_version="stable"
)
print(f"AnnData shape: {adata.shape}")
print(f"Genes: {adata.n_vars}")
print(f"Cells: {adata.n_obs}")

# Step 4: Basic QC and filtering with scanpy
print("\n3. Performing quality control...")
sc.pp.filter_cells(adata, min_genes=200)
sc.pp.filter_genes(adata, min_cells=3)
print(f"After QC - Cells: {adata.n_obs}, Genes: {adata.n_vars}")

# Step 5: Normalize and log-transform
print("\n4. Normalizing data...")
sc.pp.normalize_total(adata, target_sum=1e4)
sc.pp.log1p(adata)

# Step 6: Calculate gene expression statistics
print("\n5. Calculating expression statistics...")
for gene in genes_of_interest:
    if gene in adata.var_names:
        expr = adata[:, gene].X.toarray().flatten()
        print(f"\n{gene} expression:")
        print(f"  Mean: {expr.mean():.3f}")
        print(f"  Median: {np.median(expr):.3f}")
        print(f"  % expressing: {(expr > 0).sum() / len(expr) * 100:.1f}%")

# Step 7: Get tissue expression from ARCHS4 for comparison
print("\n6. Getting bulk tissue expression from ARCHS4...")
for gene in genes_of_interest:
    tissue_expr = gget.archs4(gene, which="tissue")
    lung_expr = tissue_expr[tissue_expr["tissue"] == "lung"]
    if len(lung_expr) > 0:
        print(f"\n{gene} in lung (ARCHS4):")
        print(f"  Median: {lung_expr['median'].iloc[0]:.3f}")

# Step 8: Enrichment analysis
print("\n7. Performing enrichment analysis...")
enrichment = gget.enrichr(genes_of_interest, database="celltypes", plot=True)
print("\nTop cell type associations:")
print(enrichment.head(10))

# Step 9: Get disease associations
print("\n8. Getting disease associations...")
for gene in genes_of_interest:
    gene_search = gget.search([gene], species="homo_sapiens", limit=1)
    if len(gene_search) > 0:
        gene_id = gene_search["ensembl_id"].iloc[0]
        diseases = gget.opentargets(gene_id, resource="diseases", limit=3)
        print(f"\n{gene} disease associations:")
        print(diseases[["disease_name", "overall_score"]])

print("\nSingle-cell expression analysis completed!")

---

Building Reference Transcriptomes

Prepare reference data for RNA-seq analysis pipelines.

#!/bin/bash
# Reference transcriptome building workflow

echo "Reference Transcriptome Building Workflow"
echo "=========================================="

# Step 1: List available species
echo -e "\n1. Listing available species..."
gget ref --list_species > available_species.txt
echo "Available species saved to available_species.txt"

# Step 2: Download reference files for human
echo -e "\n2. Downloading human reference files..."
SPECIES="homo_sapiens"
RELEASE=110  # Specify release for reproducibility

# Download GTF annotation
echo "Downloading GTF annotation..."
gget ref -w gtf -r $RELEASE -d $SPECIES -o human_ref_gtf.json

# Download cDNA sequences
echo "Downloading cDNA sequences..."
gget ref -w cdna -r $RELEASE -d $SPECIES -o human_ref_cdna.json

# Download protein sequences
echo "Downloading protein sequences..."
gget ref -w pep -r $RELEASE -d $SPECIES -o human_ref_pep.json

# Step 3: Build kallisto index (if kallisto is installed)
echo -e "\n3. Building kallisto index..."
if command -v kallisto &> /dev/null; then
    # Get cDNA FASTA file from download
    CDNA_FILE=$(ls *.cdna.all.fa.gz)
    if [ -f "$CDNA_FILE" ]; then
        kallisto index -i transcriptome.idx $CDNA_FILE
        echo "Kallisto index created: transcriptome.idx"
    else
        echo "cDNA FASTA file not found"
    fi
else
    echo "kallisto not installed, skipping index building"
fi

# Step 4: Download genome for alignment-based methods
echo -e "\n4. Downloading genome sequence..."
gget ref -w dna -r $RELEASE -d $SPECIES -o human_ref_dna.json

# Step 5: Get gene information for genes of interest
echo -e "\n5. Getting information for specific genes..."
gget search -s $SPECIES "TP53 BRCA1 BRCA2" -o key_genes.csv

echo -e "\nReference transcriptome building completed!"

# Python version
import gget
import json

print("Reference Transcriptome Building Workflow")
print("=" * 50)

# Configuration
species = "homo_sapiens"
release = 110
genes_of_interest = ["TP53", "BRCA1", "BRCA2", "MYC", "EGFR"]

# Step 1: Get reference information
print("\n1. Getting reference information...")
ref_info = gget.ref(species, release=release)

# Save reference information
with open("reference_info.json", "w") as f:
    json.dump(ref_info, f, indent=2)
print("Reference information saved to reference_info.json")

# Step 2: Download specific files
print("\n2. Downloading reference files...")
# GTF annotation
gget.ref(species, which="gtf", release=release, download=True)
# cDNA sequences
gget.ref(species, which="cdna", release=release, download=True)

# Step 3: Get information for genes of interest
print(f"\n3. Getting information for {len(genes_of_interest)} genes...")
gene_data = []
for gene in genes_of_interest:
    result = gget.search([gene], species=species, limit=1)
    if len(result) > 0:
        gene_data.append(result.iloc[0])

# Get detailed info
if gene_data:
    gene_ids = [g["ensembl_id"] for g in gene_data]
    detailed_info = gget.info(gene_ids)
    detailed_info.to_csv("genes_of_interest_info.csv", index=False)
    print("Gene information saved to genes_of_interest_info.csv")

# Step 4: Get sequences
print("\n4. Retrieving sequences...")
sequences_nt = gget.seq(gene_ids)
sequences_aa = gget.seq(gene_ids, translate=True)

with open("key_genes_nucleotide.fasta", "w") as f:
    f.write(sequences_nt)
with open("key_genes_protein.fasta", "w") as f:
    f.write(sequences_aa)

print("\nReference transcriptome building completed!")
print(f"Files created:")
print("  - reference_info.json")
print("  - genes_of_interest_info.csv")
print("  - key_genes_nucleotide.fasta")
print("  - key_genes_protein.fasta")

---

Mutation Impact Assessment

Analyze the impact of genetic mutations on protein structure and function.

import gget
import pandas as pd

print("Mutation Impact Assessment Workflow")
print("=" * 50)

# Define mutations to analyze
mutations = [
    {"gene": "TP53", "mutation": "c.818G>A", "description": "R273H hotspot"},
    {"gene": "EGFR", "mutation": "c.2573T>G", "description": "L858R activating"},
]

# Step 1: Get gene information
print("\n1. Getting gene information...")
for mut in mutations:
    results = gget.search([mut["gene"]], species="homo_sapiens", limit=1)
    if len(results) > 0:
        mut["ensembl_id"] = results["ensembl_id"].iloc[0]
        print(f"{mut['gene']}: {mut['ensembl_id']}")

# Step 2: Get sequences
print("\n2. Retrieving wild-type sequences...")
for mut in mutations:
    # Get nucleotide sequence
    nt_seq = gget.seq(mut["ensembl_id"])
    mut["wt_sequence"] = nt_seq

    # Get protein sequence
    aa_seq = gget.seq(mut["ensembl_id"], translate=True)
    mut["wt_protein"] = aa_seq

# Step 3: Generate mutated sequences
print("\n3. Generating mutated sequences...")
# Create mutation dataframe for gget mutate
mut_df = pd.DataFrame({
    "seq_ID": [m["gene"] for m in mutations],
    "mutation": [m["mutation"] for m in mutations]
})

# For each mutation
for mut in mutations:
    # Extract sequence from FASTA
    lines = mut["wt_sequence"].split("\n")
    seq = "".join(lines[1:])

    # Create single mutation df
    single_mut = pd.DataFrame({
        "seq_ID": [mut["gene"]],
        "mutation": [mut["mutation"]]
    })

    # Generate mutated sequence
    mutated = gget.mutate([seq], mutations=single_mut)
    mut["mutated_sequence"] = mutated

print("Mutated sequences generated")

# Step 4: Get existing structure information
print("\n4. Getting structure information...")
for mut in mutations:
    # Get info with PDB IDs
    info = gget.info([mut["ensembl_id"]], pdb=True)

    if "pdb_id" in info.columns and pd.notna(info["pdb_id"].iloc[0]):
        pdb_ids = info["pdb_id"].iloc[0].split(";")
        print(f"\n{mut['gene']} PDB structures: {', '.join(pdb_ids[:3])}")

        # Download first structure
        if len(pdb_ids) > 0:
            pdb_id = pdb_ids[0].strip()
            mut["pdb_id"] = pdb_id
            gget.pdb(pdb_id, save=True)
    else:
        print(f"\n{mut['gene']}: No PDB structure available")
        mut["pdb_id"] = None

# Step 5: Predict structures with AlphaFold (optional)
print("\n5. Predicting structures with AlphaFold...")
# Note: Requires gget setup alphafold and is computationally intensive
# Uncomment to run:
# for mut in mutations:
#     print(f"Predicting {mut['gene']} wild-type structure...")
#     wt_structure = gget.alphafold(mut["wt_protein"])
#
#     print(f"Predicting {mut['gene']} mutant structure...")
#     # Would need to translate mutated sequence first
#     # mutant_structure = gget.alphafold(mutated_protein)
print("(AlphaFold prediction skipped - uncomment to run)")

# Step 6: Find functional motifs
print("\n6. Identifying functional motifs...")
# Note: Requires gget setup elm
# Uncomment to run:
# for mut in mutations:
#     ortholog_df, regex_df = gget.elm(mut["wt_protein"])
#     print(f"\n{mut['gene']} functional motifs:")
#     print(regex_df)
print("(ELM analysis skipped - uncomment to run)")

# Step 7: Get disease associations
print("\n7. Getting disease associations...")
for mut in mutations:
    diseases = gget.opentargets(
        mut["ensembl_id"],
        resource="diseases",
        limit=5
    )
    print(f"\n{mut['gene']} ({mut['description']}) disease associations:")
    print(diseases[["disease_name", "overall_score"]])

# Step 8: Query COSMIC for mutation frequency
print("\n8. Querying COSMIC database...")
# Note: Requires COSMIC database download
# Uncomment to run:
# for mut in mutations:
#     cosmic_results = gget.cosmic(
#         mut["mutation"],
#         cosmic_tsv_path="cosmic_cancer.tsv",
#         limit=10
#     )
#     print(f"\n{mut['gene']} {mut['mutation']} in COSMIC:")
#     print(cosmic_results)
print("(COSMIC query skipped - requires database download)")

print("\nMutation impact assessment completed!")

---

Drug Target Discovery

Identify and validate potential drug targets for specific diseases.

import gget
import pandas as pd

print("Drug Target Discovery Workflow")
print("=" * 50)

# Step 1: Search for disease-related genes
disease = "alzheimer"
print(f"\n1. Searching for {disease} disease genes...")
genes = gget.search([disease], species="homo_sapiens", limit=50)
print(f"Found {len(genes)} potential genes")

# Step 2: Get detailed information
print("\n2. Getting detailed gene information...")
gene_ids = genes["ensembl_id"].tolist()[:20]  # Top 20
gene_info = gget.info(gene_ids[:10])  # Limit to avoid timeout

# Step 3: Get disease associations from OpenTargets
print("\n3. Getting disease associations...")
disease_scores = []
for gene_id, gene_name in zip(gene_info["ensembl_id"], gene_info["gene_name"]):
    diseases = gget.opentargets(gene_id, resource="diseases", limit=10)

    # Filter for Alzheimer's disease
    alzheimer = diseases[diseases["disease_name"].str.contains("Alzheimer", case=False, na=False)]

    if len(alzheimer) > 0:
        disease_scores.append({
            "ensembl_id": gene_id,
            "gene_name": gene_name,
            "disease_score": alzheimer["overall_score"].max()
        })

disease_df = pd.DataFrame(disease_scores).sort_values("disease_score", ascending=False)
print("\nTop disease-associated genes:")
print(disease_df.head(10))

# Step 4: Get tractability information
print("\n4. Assessing target tractability...")
top_targets = disease_df.head(5)
for _, row in top_targets.iterrows():
    tractability = gget.opentargets(
        row["ensembl_id"],
        resource="tractability"
    )
    print(f"\n{row['gene_name']} tractability:")
    print(tractability)

# Step 5: Get expression data
print("\n5. Getting tissue expression data...")
for _, row in top_targets.iterrows():
    # Brain expression from OpenTargets
    expression = gget.opentargets(
        row["ensembl_id"],
        resource="expression",
        filter_tissue="brain"
    )
    print(f"\n{row['gene_name']} brain expression:")
    print(expression)

    # Tissue expression from ARCHS4
    tissue_expr = gget.archs4(row["gene_name"], which="tissue")
    brain_expr = tissue_expr[tissue_expr["tissue"].str.contains("brain", case=False, na=False)]
    print(f"ARCHS4 brain expression:")
    print(brain_expr)

# Step 6: Check for existing drugs
print("\n6. Checking for existing drugs...")
for _, row in top_targets.iterrows():
    drugs = gget.opentargets(row["ensembl_id"], resource="drugs", limit=5)
    print(f"\n{row['gene_name']} drug associations:")
    if len(drugs) > 0:
        print(drugs[["drug_name", "drug_type", "max_phase_for_all_diseases"]])
    else:
        print("No drugs found")

# Step 7: Get protein-protein interactions
print("\n7. Getting protein-protein interactions...")
for _, row in top_targets.iterrows():
    interactions = gget.opentargets(
        row["ensembl_id"],
        resource="interactions",
        limit=10
    )
    print(f"\n{row['gene_name']} interacts with:")
    if len(interactions) > 0:
        print(interactions[["gene_b_symbol", "interaction_score"]])

# Step 8: Enrichment analysis
print("\n8. Performing pathway enrichment...")
gene_list = top_targets["gene_name"].tolist()
enrichment = gget.enrichr(gene_list, database="pathway", plot=True)
print("\nTop enriched pathways:")
print(enrichment.head(10))

# Step 9: Get structure information
print("\n9. Getting structure information...")
for _, row in top_targets.iterrows():
    info = gget.info([row["ensembl_id"]], pdb=True)

    if "pdb_id" in info.columns and pd.notna(info["pdb_id"].iloc[0]):
        pdb_ids = info["pdb_id"].iloc[0].split(";")
        print(f"\n{row['gene_name']} PDB structures: {', '.join(pdb_ids[:3])}")
    else:
        print(f"\n{row['gene_name']}: No PDB structure available")
        # Could predict with AlphaFold
        print(f"  Consider AlphaFold prediction")

# Step 10: Generate target summary report
print("\n10. Generating target summary report...")
report = []
for _, row in top_targets.iterrows():
    report.append({
        "Gene": row["gene_name"],
        "Ensembl ID": row["ensembl_id"],
        "Disease Score": row["disease_score"],
        "Target Status": "High Priority"
    })

report_df = pd.DataFrame(report)
report_df.to_csv("drug_targets_report.csv", index=False)
print("\nTarget report saved to drug_targets_report.csv")

print("\nDrug target discovery workflow completed!")

---

Tips for Workflow Development

Error Handling

import gget

def safe_gget_call(func, *args, **kwargs):
    """Wrapper for gget calls with error handling"""
    try:
        result = func(*args, **kwargs)
        return result
    except Exception as e:
        print(f"Error in {func.__name__}: {str(e)}")
        return None

# Usage
result = safe_gget_call(gget.search, ["ACE2"], species="homo_sapiens")
if result is not None:
    print(result)

Rate Limiting

import time
import gget

def rate_limited_queries(gene_ids, delay=1):
    """Query multiple genes with rate limiting"""
    results = []
    for i, gene_id in enumerate(gene_ids):
        print(f"Querying {i+1}/{len(gene_ids)}: {gene_id}")
        result = gget.info([gene_id])
        results.append(result)

        if i < len(gene_ids) - 1:  # Don't sleep after last query
            time.sleep(delay)

    return pd.concat(results, ignore_index=True)

Caching Results

import os
import pickle
import gget

def cached_gget(cache_file, func, *args, **kwargs):
    """Cache gget results to avoid repeated queries"""
    if os.path.exists(cache_file):
        print(f"Loading from cache: {cache_file}")
        with open(cache_file, "rb") as f:
            return pickle.load(f)

    result = func(*args, **kwargs)

    with open(cache_file, "wb") as f:
        pickle.dump(result, f)
    print(f"Saved to cache: {cache_file}")

    return result

# Usage
result = cached_gget("ace2_info.pkl", gget.info, ["ENSG00000130234"])

---

These workflows demonstrate how to combine multiple gget modules for comprehensive bioinformatics analyses. Adapt them to your specific research questions and data types.