Alphagenome Single Variant Analysis

Google DeepMindOfficialPublished by Google DeepMind

Skill
Tool integration
Science and research
Apache-2.0
Low risk
Has scripts

From the author

Analyzes genetic variant effects on gene expression (RNA-seq), chromatin accessibility (DNASE), histone marks (ChIP), and transcription factors using the AlphaGenome API. Use when the user asks about non-coding variant effects, pathogenicity, clinical significance, disease associations, functional effects, gene expression changes, splicing disruption, or regulatory effects in promoters and enhancers. Also use for resolving biological terms to tissue/cell-type ontologies (UBERON/CL) or analyzing variants in chr:pos:ref>alt format.

— the description by Google DeepMind

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

The AlphaGenome Single Variant Analysis skill provides a structured workflow for investigating how specific genetic variants influence biological processes. It leverages the AlphaGenome API to predict effects on RNA-seq expression, chromatin accessibility, histone marks, and transcription factor binding. The skill includes robust procedures for resolving biological ontologies, performing discovery scans for tissue-specific effects, and visualizing variant impacts through detailed genomic tracks and splicing analysis. It emphasizes rigorous interpretation through standardized reporting templates and mandatory adherence to provided interpretation guides.

What it does

Performs discovery scans to identify significant variant effects across diverse tissues and modalities.
Resolves biological terms and tissue/cell-type names to standard UBERON/CL ontologies.
Visualizes variant effects on gene expression, splicing, and regulatory regions.
Generates structured, evidence-based analysis reports following predefined templates.
Supports in-silico mutagenesis (ISM) and quantitative splicing analysis.
Provides tools for local gene and transcript lookup using GTF data.

When to use it

When investigating the functional consequences of non-coding genetic variants.
When assessing the pathogenicity or clinical significance of a variant in a specific disease context.
When analyzing how a variant disrupts regulatory elements like promoters and enhancers.
When exploring tissue-specific gene expression changes or splicing disruption caused by a variant.
When you need to map biological terms to standardized ontologies for downstream analysis.

Advantages

Provides a standardized, reproducible workflow for complex genomic variant interpretation.
Integrates multi-omic data types (RNA-seq, DNASE, ChIP) for comprehensive functional insights.
Includes specific strategies for handling complex events like splicing skipping and large genes.
Enforces rigorous reporting standards to ensure consistency and clarity in findings.
Utilizes local GTF data for efficient and offline-capable gene lookups.

Usage examples

“Analyze the functional impact of the variant chr2:1234:A>C in liver tissue.”: The agent will perform a discovery scan, filter for liver-related biosamples, generate visualization plots for the variant, and produce a structured report detailing the predicted regulatory and expression effects.
“What is the effect of chr1:5000:G>T on splicing in neurons?”: The agent will use the interpret_splicing script to quantify delta scores and junction changes, visualize the splicing event with appropriate flanking exons, and summarize the findings in a report.

Install

Download the skill as a ZIP:
https://autogineer.com/agent-skills/google-deepmind/alphagenome-single-variant-analysis/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip alphagenome-single-variant-analysis.zip -d .agents/skills

Documentation by Google DeepMind

Variant Analysis using AlphaGenome

Prerequisites

uv: Read the uv skill and follow its Setup instructions to ensure uv is installed and on PATH.
User Notification: If .licenses/alphagenome_single_variant_analysis_LICENSE.txt does not already exist in the workspace root directory then (1) prominently notify the user to check the terms at https://deepmind.google.com/science/alphagenome/, then (2) create the file recording the notification text and timestamp.
.env file: Make sure the .env file exists in your home directory. Create one if it does not exist.
ALPHAGENOME_API_KEY: This skill requires an API key to function.
You can register for a key at https://deepmind.google.com/science/alphagenome/. You MUST use the safe credentials protocol in the credentials skill to check for and request this key if this skill looks relevant to the user's request.
ALPHAGENOME_GTF_PATH (Optional): Accelerate gene/transcript lookup by pointing to a local copy of the GTF feather file instead of downloading from GCS:
```
echo "ALPHAGENOME_GTF_PATH=/path/to/local/gencode.v46.annotation.gtf.gz.feather" >> ~/.env
```

Core Rules

NEVER run python3 or python3 -c directly. The system Python does not necessarily have pandas, numpy, and other key dependencies. ALWAYS use uv run to run ALL Python code — including scripts, ad-hoc analysis files, and one-liners. Do not attempt to pip install or create new venvs — uv manages an isolated environment automatically.
Offline Only: NEVER use external APIs (e.g., MyGene.info, Ensembl REST) for gene/transcript lookup. Use lookup_gene_info.py with the local GTF. If it fails, fix the environment/paths, do not switch to external APIs.
API Key is required: ALPHAGENOME_API_KEY must be set before running any script.
Notification: If this skill is used, ensure this is mentioned in the output.
Report Format: Always use the templates in docs/report-templates.md for generating analysis reports, and ensure to include the table of top hits from the discovery scan.

Environment Setup & Troubleshooting

Python Environment

All scripts must be executed using uv run, which manages an isolated virtual environment with the correct dependencies via uv.

uv run <script_name> [args...]

For ad-hoc scripts (e.g., inline analysis code saved to a temp file), pass the full path instead of a short name:

uv run --project $SKILL_DIR /tmp/my_analysis.py --arg1 val1

[!NOTE] The first invocation resolves and installs dependencies (~10s). Subsequent runs use the cached environment and start instantly. The cache lives in ~/.cache/uv/.

Common Issues

Column Names: tidy_scores and metadata often use gene_name (not gene_symbol) and output_type (not modality). Always inspect df.columns before filtering.
Large Genes: Genes > 500kb (e.g., USH2A) break the whole_gene view. Use --view detail or manual regional windows instead.
Sashimi Strand Error: plot_components.Sashimi does NOT accept a strand argument directly. Filter input tracks instead.
KeyError: 'ontology_curie': Not all tracks have ontology_curie. Check track.metadata.columns before filtering.
Python Path: If exec: "python": executable file not found occurs, ensure you are using uv run instead of bare python/python3.
NotImplementedError (pandas): "iLocation based boolean indexing on an integer type is not available". This occurs when using boolean masks with .iloc on integer-indexed DataFrames in newer pandas versions. Fix: Convert boolean masks to integer indices using np.flatnonzero(mask).
GTF Feather Case Sensitivity: The AlphaGenome GTF Feather file uses Capitalized column names (Feature, Start, End, Strand) unlike standard GTF files. Always check df.columns if getting KeyErrors.
score_variant ontology filtering: score_variant does NOT accept ontology_terms as an argument. You must filter the returned AnnData objects manually by inspecting adata.var columns. In contrast, predict_variant DOES accept ontology_terms directly.
Sashimi Zoom Logic: To ensure "skipping" arcs are visible, expand the zoom to include the flanking exons rather than relying on junction overlap alone.
Junction Scores: Raw Junction objects from prediction may be simple Intervals. Use junction_data.get_junctions_to_plot(predictions=..., name=...) to retrieve objects with the .k (abundance/score) attribute.
uv Not Found: If exec: uv: not found, follow the installation instructions in Prerequisites.
Registry Authentication Error (401): If uv fails with 401 Unauthorized for a private registry, set UV_INDEX_URL=https://pypi.org/simple before running the script.

References

alphagenome-api.md — API reference and code patterns
interpretation-guide.md — Interpretation guide, score magnitude rules, ISM, and checklist.
report-templates.md — Full report templates
scripts/visualize_variant_effects.py — Single-variant visualization template (Ref/Alt comparisons, Splicing).
- Splicing Zoom Strategy: Uses a Hybrid Approach for optimal visibility:
  1. Base Interval: Variant +/- 1 downstream and upstream exon (Structural Context).
  2. Junction Expansion: Expands to include the full span of any significant splicing junction (e.g., exon skipping events that span multiple exons).
  3. Anchor Enforcement: Ensures the exons anchoring these long junctions are fully visible. Lesson: Simple fixed windows (e.g., 2kb) or nearest-exon logic often fail for skipping events. Always use the observed junction data to drive zoom levels.
examples/splicing/ — Splicing analysis examples
examples/model_limitation_RNU4ATAC/ — ncRNA structure limitation case study
examples/polyadenylation_HBA2/ — 3' UTR / Polyadenylation case study
examples/regulatory/ — Regulatory variant examples
examples/negative_result_GATA4/ — Negative results (mathematical artefact)
examples/negative_result_TGFB3/ — Negative results (proxies)
scripts/lookup_gene_info.py — Gene & transcript lookup
scripts/resolve_ontology_terms.py — Ontology term resolution (UBERON/CL IDs)

Code Patterns

Broad Discovery Scan

Use score_variant across differential scorers only to discover unexpected tissue effects.

from alphagenome.models import dna_client
from alphagenome.models import variant_scorers
from alphagenome.data import genome
import os
import pandas as pd
import dotenv

# Load environment variables from ~/.env
dotenv.load_dotenv(os.path.expanduser('~/.env'))

# Setup API Key and Client
dna_model = dna_client.create(api_key=os.environ.get('ALPHAGENOME_API_KEY'),
                              address='dns:///gdmscience.googleapis.com:443')

# Define Variant (example)
variant_str = "chr2:1234:A>C"
chrom, pos_str, ref_alt = variant_str.split(':')
ref, alt = ref_alt.split('>')
pos = int(pos_str)

# Use supported sequence length (e.g., 2**20 for optimal performance)
SEQ_LENGTH = 2**20
interval = genome.Interval(chrom, pos - SEQ_LENGTH // 2, pos + SEQ_LENGTH // 2)
variant = genome.Variant(chrom, pos, ref, alt)

scorers = [
    variant_scorers.RECOMMENDED_VARIANT_SCORERS[m]
    for m in variant_scorers.RECOMMENDED_VARIANT_SCORERS
    if "ACTIVE" not in m and "CAGE" not in m and "PROCAP" not in m
]

print(f"Scoring variant {variant_str}...")
scores_list = dna_model.score_variant(interval=interval, variant=variant, variant_scorers=scorers)

# Process and Display Results
all_dfs = []
for score_adata in scores_list:
    df = variant_scorers.tidy_scores([score_adata], match_gene_strand=True)
    if df is not None:
        all_dfs.append(df)

if all_dfs:
    df = pd.concat(all_dfs)
    significant = df[df['quantile_score'].abs() > 0.995]
    ranked = significant.sort_values('raw_score', key=abs, ascending=False)
    print("Top Significant Hits:")
    print(ranked[['biosample_name', 'gene_name', 'output_type', 'quantile_score', 'raw_score']])

Extended Search for Disease-Relevant Tissues

# Define keywords based on disease context
disease_keywords = ["liver", "hepatocyte"]

# Filter for any match
mask = df['biosample_name'].str.contains('|'.join(disease_keywords), case=False, na=False)

relevant_hits = df[mask].sort_values('raw_score', key=abs, ascending=False)
print(f"\n--- Extended Analysis (Keywords: {disease_keywords}) ---")
print(relevant_hits.head(20)[['biosample_name', 'output_type', 'raw_score', 'quantile_score']])

Workflow Checklist

Variant Analysis Progress:
- [ ] Step 0: Review Golden Examples (MANDATORY)
- [ ] Step 1: Create Output Folder and Setup
- [ ] Step 2: Parse User Query & Research
- [ ] Step 3: Resolve Tissues & Modalities
- [ ] Step 4: Visualize & Save Plots
- [ ] Step 5: Analyze Predictions (view plots, no code). MANDATORY: Read [interpretation-guide.md](docs/interpretation-guide.md) before interpreting results.
- [ ] Step 6: Write Report, save it as `report.md` (MANDATORY)
- [ ] Step 7: Self-Critique (view `report.md` to verify links & claims)
- [ ] Step 8: Make artifact out of `report.md`

Multi-Variant Workflow

If multiple variants are specified, spawn sub-agents to run each variant analysis and then synthesize each report.md into a single report.

Script Reference

Script	Purpose
`lookup_gene_info`	Comprehensive gene and transcript lookup using
: : GTF data :
`resolve_ontology_terms`	Biological terms → UBERON/CL/EFO IDs
`visualize_variant_effects`	REF/ALT visualization (expression, regulatory,
: : splicing) :
`analyze_ism`	In-Silico Mutagenesis SeqLogo generation
`interpret_splicing`	Quantitative splicing analysis (delta scores,
: : junctions) :
`visualize_genome_tracks`	Genomic track visualization for a region

Files

ATTRIBUTION.mdAdded by AutoGineer952 B
LICENSELicence file added by AutoGineer11.1 KiB
SKILL.md11.7 KiB
docs/alphagenome-api.md13.3 KiB
docs/examples/model_limitation_RNU4ATAC/README.md1.7 KiB
docs/examples/model_limitation_RNU4ATAC/ism_skeletal_muscle_RNA_SEQ.png12.1 KiB
docs/examples/model_limitation_RNU4ATAC/plot_skeletal_muscle_RNU4ATAC_detail.png76.3 KiB
docs/examples/model_limitation_RNU4ATAC/plot_skeletal_muscle_RNU4ATAC_effects.png72.9 KiB
docs/examples/model_limitation_RNU4ATAC/plot_skeletal_muscle_RNU4ATAC_wholegene.png78.3 KiB
docs/examples/model_limitation_RNU4ATAC/report.md3.7 KiB
docs/examples/negative_result_GATA4/README.md1.1 KiB
docs/examples/negative_result_GATA4/ism_heart_DNASE.png11.7 KiB
docs/examples/negative_result_GATA4/plot_heart_GATA4_detail.png73.5 KiB
docs/examples/negative_result_GATA4/plot_heart_GATA4_effects.png70.2 KiB
docs/examples/negative_result_GATA4/plot_heart_GATA4_wholegene.png78.2 KiB
docs/examples/negative_result_GATA4/report.md3.6 KiB
docs/examples/negative_result_TGFB3/README.md1.2 KiB
docs/examples/negative_result_TGFB3/ism_heart_ATAC.png6.5 KiB
docs/examples/negative_result_TGFB3/ism_heart_RNA_SEQ.png10.3 KiB
docs/examples/negative_result_TGFB3/plot_heart_TGFB3_detail.png51.5 KiB
docs/examples/negative_result_TGFB3/plot_heart_TGFB3_effects.png64.0 KiB
docs/examples/negative_result_TGFB3/plot_heart_TGFB3_wholegene.png43.9 KiB
docs/examples/negative_result_TGFB3/report.md3.0 KiB
docs/examples/polyadenylation_HBA2/README.md1.1 KiB
docs/examples/polyadenylation_HBA2/ism_K562_RNA_SEQ.png10.8 KiB
docs/examples/polyadenylation_HBA2/plot_K562_HBA2_effects.png158.5 KiB
docs/examples/polyadenylation_HBA2/plot_K562_HBA2_wholegene.png159.3 KiB
docs/examples/polyadenylation_HBA2/report.md3.1 KiB
docs/examples/regulatory/README.md1.2 KiB
docs/examples/regulatory/apoa1_promoter/ism_heart_left_ventricle_RNA_SEQ.png8.6 KiB
docs/examples/regulatory/apoa1_promoter/ism_liver_RNA_SEQ.png8.6 KiB
docs/examples/regulatory/apoa1_promoter/plot_heart_left_ventricle_APOA1_effects.png73.0 KiB
docs/examples/regulatory/apoa1_promoter/plot_liver_APOA1_effects.png221.5 KiB
docs/examples/regulatory/apoa1_promoter/report.md3.4 KiB
docs/examples/splicing/README.md1.3 KiB
docs/examples/splicing/col6a2_exon_extension.png101.4 KiB
docs/examples/splicing/col6a2_report.md2.2 KiB
docs/examples/splicing/dlg1_exon_skipping.png73.1 KiB
docs/examples/splicing/dlg1_report.md1.9 KiB
docs/interpretation-guide.md17.4 KiB
docs/report-templates.md2.8 KiB
pyproject.toml269 B
references/citation.bib823 B
scripts/analyze_ism.pyScript10.1 KiB
scripts/generate_ontology_mapping.pyScript4.8 KiB
scripts/interpret_splicing.pyScript9.5 KiB
scripts/lookup_gene_info.pyScript9.8 KiB
scripts/resolve_ontology_terms.pyScript5.5 KiB
scripts/visualize_genome_tracks.pyScript12.4 KiB
scripts/visualize_variant_effects.pyScript37.0 KiB

docs/alphagenome-api.md

AlphaGenome API Reference

Pip package: alphagenome

Setup and Imports

Standard imports for AlphaGenome workflows:

from alphagenome.data import gene_annotation
from alphagenome.data import genome
from alphagenome.data import track_data
from alphagenome.data import transcript as transcript_utils
from alphagenome.interpretation import ism
from alphagenome.models import dna_client
from alphagenome.models import variant_scorers
from alphagenome.visualization import plot_components
import matplotlib.pyplot as plt
import pandas as pd

Client Initialization

The API key is automatically loaded by dotenv from the .env file in the agent configuration dir.

To initialize a client:

from alphagenome.models import dna_client

dna_model = dna_client.create(
    api_key=os.environ.get('ALPHAGENOME_API_KEY'),
    address='dns:///gdmscience.googleapis.com:443',
)

Core Data Types

genome.Interval

0-based half-open interval (includes start, excludes end).

interval = genome.Interval(chromosome='chr1', start=1_000, end=1_010)

interval.center()     # Returns center position (int)
interval.width        # Returns 10
interval.resize(100)  # Resizes around center
interval.overlaps(other_interval)
interval.contains(other_interval)
interval.intersect(other_interval)

genome.Variant

Position is 1-based (VCF-compatible).

variant = genome.Variant(
    chromosome='chr22',
    position=36201698,  # 1-based!
    reference_bases='A',
    alternate_bases='C',
)
# Get interval around variant
interval = variant.reference_interval.resize(dna_client.SEQUENCE_LENGTH_1MB)

Predictions

Predict from DNA Sequence

output = dna_model.predict_sequence(
    sequence='GATTACA'.center(dna_client.SEQUENCE_LENGTH_1MB, 'N'),
    requested_outputs=[dna_client.OutputType.DNASE],
    ontology_terms=['UBERON:0002048'],  # Lung
)
# Access predictions
print(output.dnase.values.shape)  # (sequence_length, num_tracks)
print(output.dnase.metadata)      # Track metadata DataFrame

Predict from Genome Interval

interval = genome.Interval('chr1', 1000000, 1000001)
interval = interval.resize(dna_client.SEQUENCE_LENGTH_1MB)

output = dna_model.predict_interval(
    interval=interval,
    requested_outputs=[dna_client.OutputType.RNA_SEQ],
    ontology_terms=['UBERON:0001114'],  # Right liver lobe
)

Mouse Predictions

Specify organism=dna_client.Organism.MUS_MUSCULUS for mouse models.

output = dna_model.predict_sequence(
    ...,
    organism=dna_client.Organism.MUS_MUSCULUS,
)

Variant Analysis

Predict Variant Effects (Raw Tracks)

Compare predictions for Reference (REF) vs Alternate (ALT) alleles.

variant_output = dna_model.predict_variant(
    interval=interval,
    variant=variant,
    requested_outputs=[dna_client.OutputType.RNA_SEQ],
    ontology_terms=['UBERON:0001157'],  # Colon - Transverse
)

ref_tracks = variant_output.reference.rna_seq
alt_tracks = variant_output.alternate.rna_seq

Score Variants (Aggregated Scores)

Get aggregated scores using recommended scorers.

scorer = variant_scorers.RECOMMENDED_VARIANT_SCORERS['RNA_SEQ']

variant_scores_list = dna_model.score_variant(
    interval=interval,
    variant=variant,
    variant_scorers=[scorer],
)
scores = variant_scores_list[0]

# Tidy scores to DataFrame
df = variant_scorers.tidy_scores([scores], match_gene_strand=True)
print(df[['gene_symbol', 'raw_score', 'quantile_score']])

Available recommended scorers: ATAC, CAGE, DNASE, PROCAP, RNA_SEQ, CHIP_TF, CHIP_HISTONE, SPLICE_SITES, SPLICE_SITE_USAGE, SPLICE_JUNCTIONS, POLYADENYLATION, CONTACT_MAPS

Batch Variant Scoring

# Parse variants from VCF-like DataFrame
for _, row in vcf_df.iterrows():
    variant = genome.Variant(
        chromosome=str(row.CHROM),
        position=int(row.POS),
        reference_bases=row.REF,
        alternate_bases=row.ALT,
    )
    interval = variant.reference_interval.resize(
        dna_client.SEQUENCE_LENGTH_1MB
    )
    scores = dna_model.score_variant(
        interval=interval,
        variant=variant,
        variant_scorers=list(
            variant_scorers.RECOMMENDED_VARIANT_SCORERS.values()
        ),
    )

In Silico Mutagenesis (ISM)

Systematically mutate a region to find important motifs.

from alphagenome.interpretation import ism

sequence_interval = genome.Interval(
    'chr20', 3_753_000, 3_753_400
).resize(dna_client.SEQUENCE_LENGTH_16KB)
ism_interval = sequence_interval.resize(256)  # Mutate center 256bp

scorer = variant_scorers.CenterMaskScorer(
    requested_output=dna_client.OutputType.DNASE,
    width=501,
    aggregation_type=variant_scorers.AggregationType.DIFF_MEAN,
)

variant_scores = dna_model.score_ism_variants(
    interval=sequence_interval,
    ism_interval=ism_interval,
    variant_scorers=[scorer],
)

TrackData Operations

Properties

tdata = output.dnase
tdata.values      # numpy array (sequence_length, num_tracks)
tdata.metadata    # pandas DataFrame with track info
tdata.resolution  # bp per position
tdata.interval    # genome.Interval

Filtering by Strand

pos_tracks = tdata.filter_to_positive_strand()
neg_tracks = tdata.filter_to_negative_strand()
unstranded = tdata.filter_to_unstranded()

Filtering by Track Name

track1_tdata = tdata.select_tracks_by_name(names='track1')

Filtering by Metadata (Manual)

mask = tracks.metadata['histone_mark'] == 'H3K27ac'
filtered_tracks = track_data.TrackData(
    values=tracks.values[:, mask],
    metadata=tracks.metadata[mask],
    resolution=tracks.resolution,
    interval=tracks.interval,
    uns=tracks.uns,
)

Slicing

# By position indices
tdata.slice_by_positions(start=2, end=4)

# By genomic interval
tdata.slice_by_interval(
    genome.Interval(chromosome='chr1', start=1_002, end=1_004)
)

Resizing

tdata.resize(width=2)  # Crop to center 2 positions
tdata.resize(width=8)  # Pad with zeros

Resolution Conversion

tdata.change_resolution(resolution=128)  # Downsample
tdata.change_resolution(resolution=1)    # Upsample

Track Metadata Reference

Modality	Key Column	Example Values
`CHIP_HISTONE`	`histone_mark`	`H3K27ac`, `H3K4me3`, `H3K27me3`
`CHIP_TF`	`target`	`CTCF`, `JUND`, `POLR2A`
`RNA_SEQ`, `CAGE`	`strand`	`+`, `-`
All	`ontology_curie`	`UBERON:0002107`, `EFO:0001187`
All	`biosample_name`	`liver`, `HepG2`

[!CAUTION] CHIP_HISTONE uses histone_mark, NOT target. target is for CHIP_TF.

Visualization

plot_components.plot(
    components=[
        plot_components.TranscriptAnnotation(transcripts),
        plot_components.Tracks(output.rna_seq),
        plot_components.OverlaidTracks(
            tdata={'REF': ref_tracks, 'ALT': alt_tracks},
            colors={'REF': 'dimgrey', 'ALT': 'red'},
        ),
    ],
    interval=interval,
    annotations=[plot_components.VariantAnnotation([variant], alpha=0.8)],
)
plt.show()

[!CAUTION] VariantAnnotation must be in annotations=, NOT components. Putting it in components causes AttributeError: 'VariantAnnotation' object has no attribute 'num_axes'.

Get Human-Readable Tissue Names

tissue = tracks.metadata[
    tracks.metadata['ontology_curie'] == ontology_id
]['biosample_name'].iloc[0]
plt.title(f"{gene_symbol} - {tissue} - {modality.name}")

Gene Annotations (GTF)

gtf = pd.read_feather(
    'https://storage.googleapis.com/alphagenome/reference/gencode/'
    'hg38/gencode.v46.annotation.gtf.gz.feather'
)

# Filter for MANE Select transcripts
gtf_transcripts = gene_annotation.filter_protein_coding(gtf)
gtf_transcripts = gene_annotation.filter_to_mane_select_transcript(
    gtf_transcripts
)

# Get stranded interval for a gene
interval = gene_annotation.get_gene_interval(gtf, gene_symbol='CYP2B6')

GTF feather uses capitalized column names for core fields:

Correct	Incorrect
`Feature`	`feature`
`Chromosome`	`seqname`
`Start`	`start`
`End`	`end`
`Strand`	`strand`

Other columns (gene_name, gene_id, gene_type) remain lowercase.

Best Practices

Interval Resizing: Always resize to a supported length before prediction. Use dna_client.SUPPORTED_SEQUENCE_LENGTHS.keys() for options.
Efficient Predictions: Always specify requested_outputs and ontology_terms to reduce compute and data transfer.
Variant Scoring: Use tidy_scores(..., match_gene_strand=True) to filter irrelevant strand matches. quantile_score allows comparison across different scorers.
ISM: Expensive (scores 3 variants per position). Use shorter context intervals (16KB) and narrower mutation regions.
Saving Figures: Use plt.savefig('plot.png', bbox_inches='tight') to prevent cut-off labels.

Common Pitfalls

OverlaidTracks has no `title` argument

# Wrong:
plot_components.OverlaidTracks(..., title="My Title")  # ERROR!

# Correct:
plot_components.plot([plot_components.OverlaidTracks(...)], interval=...)
plt.suptitle("My Title")

TrackData has no `subset()` method

Use manual filtering with boolean masks on metadata, or use select_tracks_by_name() for name-based selection.

`uv run` fails

Clear the cached environment and retry: uv cache clean && uv run <script_name>. If the issue persists, check that pyproject.toml exists in the skill root and contains alphagenome>=0.6.1 in dependencies.

Client initialization needs correct address

Always use the production address for API access: address='dns:///gdmscience.googleapis.com:443'

Output Types

Defined as dna_client.OutputType, used in requested_outputs:

OutputType	Description
`ATAC`	ATAC-seq: chromatin accessibility
`CAGE`	Cap Analysis of Gene Expression
`DNASE`	DNase I hypersensitive sites: chromatin accessibility
`RNA_SEQ`	RNA sequencing: gene expression
`CHIP_HISTONE`	ChIP-seq: histone modifications
`CHIP_TF`	ChIP-seq: transcription factor binding
`SPLICE_SITES`	Donor and acceptor splice sites
`SPLICE_SITE_USAGE`	Fraction of time each splice site is used
`SPLICE_JUNCTIONS`	Split read RNA-seq counts for each junction
`CONTACT_MAPS`	3D DNA-DNA contact probabilities
`PROCAP`	Precision Run-On sequencing and capping

Variant Scoring Details

Gene Expression (RNA-seq)

Quantifies impact on overall gene transcript abundance.

Comparison: Predicted RNA coverage between REF and ALT alleles.
Mask: Exons for a gene of interest.
Aggregation: Log-fold change: log(mean(ALT) + 0.001) - log(mean(REF) + 0.001).

Polyadenylation Site (PAS) Usage

Captures variant's impact on RNA isoform production (paQTLs).

Mask: Local 400bp windows around 3' cleavage junctions.
Aggregation: Maximum absolute log-fold change of isoform ratios (distal/proximal PAS usage).

TSS Activity (CAGE, PRO-cap)

Quantifies local changes at TSSs.

Mask: Local 501bp window centered at variant.
Aggregation: log2[(sum(ALT) + 1) / (sum(REF) + 1)].

Chromatin Accessibility (ATAC-seq, DNase-seq)

Mask: Local 501bp window centered at variant.
Aggregation: log2[(sum(ALT) + 1) / (sum(REF) + 1)].

Transcription Factor Binding (ChIP-TF)

Mask: Local 501bp window centered at variant.
Aggregation: log2[(sum(ALT) + 1) / (sum(REF) + 1)].

Histone Modifications (ChIP-Histone)

Mask: Local 2001bp window centered at variant.
Aggregation: log2[(sum(ALT) + 1) / (sum(REF) + 1)].

Splicing (Splice Sites)

Changes in class assignment probabilities (acceptor, donor) across gene body.

Aggregation: max(|ALT - REF|) across gene body.

Splicing (Splice Site Usage)

Changes in fraction of splice site usage.

Aggregation: max(|ALT - REF|) across gene body.

Splicing (Splice Junctions)

Changes in predicted RNA-seq reads spanning junctions.

Aggregation: max(|log(ALT) - log(REF)|) across splice site pairs.

3D Genome Contact (Contact Maps)

Local contact disruption.

Mask: Local 1MB window centered at variant.
Aggregation: Mean absolute difference of contact frequencies for variant-containing bin.

Active Allele Scorers

Capture absolute activity level (not REF/ALT difference): max(aggregated_signal(ALT), aggregated_signal(REF)) over masked region.

Quantile Scores

Quantile scores are empirical percentile ranks vs common variants (MAF>0.01 in gnomAD v3). A quantile of 0.99 means the score is at the 99th percentile. Maximum value is ±0.999990 (~300K variant background). For signed scorers, quantiles are linearly mapped to [-1, 1] to preserve directionality.

Practical rule: Use quantile as significance indicator; use raw_score for magnitude comparison within the same scorer.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Needs an API key, Needs Python

Security

What our scanner noticed· Low risk

Instructions that use the network
Scripts that your agent may run

This skill includes scripts that your agent may run. Read them before you use it.

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Published by

Google DeepMind

Publisher

Google DeepMind (Official)

Source

google-deepmind/science-skills · skills/alphagenome_single_variant_analysis · commit 4de6786

Licence

Apache License 2.0 (redistribution allowed)

LICENSE (added by AutoGineer from LICENSE)
ATTRIBUTION.md (written by AutoGineer)

Synced

2026-11-09

AutoGineer is not affiliated with or endorsed by Google DeepMind or Google DeepMind.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download alphagenome-single-variant-analysis.zip