Published Variants and Candidate Genes

Estimated time: 25 minutes

See how the source study narrowed its whole-genome sequencing results, then examine the 54 reported variant rows.

Candidate-gene filtering and variant ranking

The source study performed whole-genome sequencing on heart tissue from people with early-onset advanced heart failure. The researchers narrowed millions of variants through quality control, gene selection, computational ranking, and clinical evaluation:

  1. They focused on 369 genes connected to heart failure and related conditions.
  2. A ranking pipeline assigned a pathogenicity score across eight categories: variant-call quality, splicing, predicted consequence, gene intolerance, conservation, protein-effect predictions, allele frequency, and ClinVar evidence.
  3. Variants with a study score of at least 15, plus previously reported likely pathogenic or pathogenic ClinVar variants, received manual review.
  4. Clinical geneticists applied ACMG guidance and condition-specific rules.

The published candidate genetic variant table

Supplementary Table S4 is the starting point for this module. Its 54 rows cover 46 people and 25 genes.

The source study’s Supplementary Methods specify PolyPhen, SIFT, REVEL, and CADD for the protein-effect component. The paper used its score, together with previously reported ClinVar classifications, to select variants for manual review. Clinical geneticists then assigned the reported ACMG classifications.

This module uses those classifications as the published starting point. GTEx, HuBMAP, Pharos, and ProtVar results then inform hypotheses and experimental follow-up.

TipWhy add information about the genes?

The candidate genetic variant table provides the clinical starting point for experimental follow-up. GTEx reports heart-tissue expression, HuBMAP reports indexed values for ventricular cardiac myocytes, and Pharos summarizes protein knowledge, ligands, and target development. The table records the study classification for each variant.

Inspect the published variants

Load the published fields

Load the 54 published variant observations and display the first five rows.

Each row records one variant observation for one participant.

Understand the key columns

The key columns describe the variant, its reported classification and phenotype, and its source.

Column Meaning
subject_id Study participant identifier
gene_symbol Standard gene symbol associated with the variant
hgvs_c Transcript-specific coding DNA HGVS description
hgvs_p Reported protein HGVS description; a blank means none was reported
study_pathogenicity_score Prioritization score reported by the source study
study_class Study-reported classification: P, LP, VUS, or (P)
phenotype Heart-failure phenotype reported for the participant
study_comment Additional information from the published table
source_doi, source_pmid, source_table Identifiers linking each row to its source
NoteHeart-failure phenotypes in this table
Label Meaning
ACHD Adult congenital heart disease
ARVC Arrhythmogenic right ventricular cardiomyopathy
ATTR-CM Transthyretin amyloid cardiomyopathy
DCM Dilated cardiomyopathy
HCM Hypertrophic cardiomyopathy
HCM* HCM phenocopy, a condition that resembles HCM
ICM Ischemic cardiomyopathy
Myocarditis Inflammation of the heart muscle

The source article groups HCM and HCM phenocopies together. This teaching table uses separate HCM, HCM*, and ATTR-CM labels from the published variant table so learners can see how those rows were recorded.

NoteReading the HGVS columns

HGVS descriptions name a variant relative to a specific reference sequence.

  • NM_001276345.2:c.776A>C
    • NM_001276345.2 is the RefSeq transcript accession and version.
    • c. indicates coding DNA coordinates.
    • 776A>C means that the reference A at coding-DNA position 776 is replaced by C.
  • p.Asp259Ala
    • p. indicates a protein-level description.
    • Asp259Ala means that aspartic acid at amino-acid position 259 is replaced by alanine.

The transcript accession and version matter because coordinates and predicted consequences can differ among transcripts and reference-sequence versions.

The HGVS columns contain the descriptions reported by the study. A blank hgvs_p means that the source left the protein HGVS field empty.

See the HGVS reference-sequence recommendations for additional guidance.

Validate the table

Before interpreting the data, confirm that the expected table was loaded. It should contain the required columns, 54 rows, and PMID 39910139. A failed assertion signals incomplete or incorrect input.

Summarize the dataset

Summarize the number of variant rows, people, genes, and missing protein HGVS annotations. Complete the method that counts distinct gene symbols.

Use the pandas method that counts distinct values in a Series.

# Count rows, identifiers, and missing HGVSp values.
dataset_overview = pd.Series(
    {
        "variant rows": len(variants),
        "subjects": variants["subject_id"].nunique(),
        "genes": variants["gene_symbol"].nunique(),
        "missing HGVSp values": variants["hgvs_p"].isna().sum(),
    },
    name="count",
).to_frame()
dataset_overview

The 54 variant rows represent 46 participants and 25 genes. Four rows have a blank protein HGVS field in the source table.

Compare classifications by phenotype

Next, count the published variant classes within each heart-failure phenotype. This comparison shows how the reported classifications are distributed across phenotypes.

TipRead the table

Which phenotype contributes the most rows? Within DCM, how many rows are classified as P, LP, and VUS?

DCM contributes 28 of the 54 rows: 11 P, 11 LP, and 6 VUS. HCM contributes 8 rows, and ARVC contributes 5.

The other 13 rows are distributed across ATTR-CM, HCM*, ICM, ACHD, and myocarditis. The table shows (P) as a separate label because the paper used it for pathogenic secondary findings.

Check your understanding

✔

Why do 54 variant rows map to only 25 genes?

Correct. A gene can have several reported variants or the same variant can occur in several participants. For example, MYBPC3 appears in multiple rows.

Participants can contribute different genes, and several participants can have observations in the same gene.

All 54 rows have a gene symbol. The smaller gene count reflects repeated genes.

✔

What does a blank hgvs_p value mean in this dataset?

Correct. The source table left the protein HGVS field blank for this row.

A blank records an empty source field. Protein-effect evidence comes from the corresponding analysis.

Key points

  • The source table contains 54 variant observations from 46 participants and 25 genes.
  • Four observations have a blank protein HGVS field in the source table.
  • The table records the study classification for each variant when later lessons add tissue expression, cell-type measurements, and protein information.

Next: Query GTEx for bulk heart-tissue expression of the 25 genes.

View Jupyter notebook