7. Variant Annotation & Prioritization

Ensembl Variant Effect Predictor (VEP) determines the effect of genomic variants (SNVs, indels, CNVs, SVs) on genes, transcripts, and protein sequences. It integrates with ClinVar, gnomAD, dbNSFP, and deep learning scores like AlphaMissense.

# 1. Annotate VCF with Ensembl VEP (ClinVar, gnomAD, HGVS, AlphaMissense)
vep \
  -i input_variants.vcf.gz \
  -o vep_annotated.vcf.gz \
  --format vcf --vcf --compress_output bgzip \
  --species homo_sapiens --assembly GRCh38 \
  --offline --cache --dir_cache ~/.vep \
  --everything \
  --af_gnomade --af_gnomadg \
  --clin_sig_allele 1 \
  --symbol --biotype --canonical \
  --fork 8

# 2. Index annotated VCF
tabix -p vcf vep_annotated.vcf.gz

SnpEff annotates variants with functional consequences (missense, nonsense, splice donor/acceptor) based on genomic coordinates and coding impact. SnpSift provides a powerful domain-specific query language to filter annotated VCF files.

# 1. Annotate variants with SnpEff
snpEff -Xmx8g GRCh38.105 input_variants.vcf.gz \
  | bgzip > snpeff_annotated.vcf.gz
tabix -p vcf snpeff_annotated.vcf.gz

# 2. Filter with SnpSift for HIGH impact mutations (stop_gained, frameshift)
SnpSift filter "(ANN[*].IMPACT = 'HIGH') & (gen[0].DP >= 20)" \
  snpeff_annotated.vcf.gz \
  | bgzip > high_impact_mutations.vcf.gz

# 3. Extract tabular TSV summary of candidate genes
SnpSift extractFields high_impact_mutations.vcf.gz \
  CHROM POS REF ALT "ANN[*].GENE" "ANN[*].EFFECT" "ANN[*].IMPACT" > candidates.tsv

ANNOVAR is a database-driven tool capable of functional annotation across gene-based, region-based, and filter-based databases, integrating OMIM disease phenotypes, COSMIC somatic mutations, and conservation metrics (GERP++, phyloP).

# 1. Convert VCF to ANNOVAR format
convert2annovar.pl -format vcf4 input_variants.vcf > sample.avinput

# 2. Run multi-database table annotation
table_annovar.pl sample.avinput humandb/ \
  -buildver hg38 \
  -out sample_annotated \
  -remove \
  -protocol refGene,cytoBand,gnomad30_genome,cosmic92,clinvar_202401,cadd16gt \
  -operation g,r,f,f,f,f \
  -nastring . \
  -csvout

OpenCRAVAT is an open-source variant interpretation platform featuring a modular plugin architecture, cancer driver scoring (CHASMplus), pathogenicity classifiers (VEST), and an interactive web GUI for clinical reporting.

# 1. Install cancer and clinical annotation modules
oc module install chasmplus vest clinvar cgc gnomad

# 2. Run OpenCRAVAT annotation on somatic VCF
oc run input_somatic.vcf \
  -l hg38 \
  -a chasmplus vest clinvar cgc \
  -t html excel \
  -d opencravat_results/

# 3. View interactive results in web browser
# oc gui opencravat_results/input_somatic.vcf.sqlite

vcfanno is a high-speed tool written in Go that annotates VCF files with any number of external VCF and BED sources based on genomic overlap, supporting custom mathematical aggregations (mean, max, concat, self).

# 1. Define TOML configuration file (conf.toml)
cat << 'EOF' > conf.toml
[[annotation]]
file = "cadd_scores.vcf.gz"
fields = ["PHRED", "RawScore"]
ops = ["self", "self"]
names = ["CADD_PHRED", "CADD_RAW"]

[[annotation]]
file = "encode_promoters.bed.gz"
columns = [4]
names = ["PROMOTER_ID"]
ops = ["concat"]
EOF

# 2. Run multi-threaded vcfanno
vcfanno -p 16 conf.toml input_variants.vcf.gz | bgzip > vcfanno_annotated.vcf.gz
tabix -p vcf vcfanno_annotated.vcf.gz

bcftools is the universal standard for manipulating variant call files in VCF and binary BCF formats. It normalizes indels against reference genomes, splits multi-allelic records, performs complex expression filtering, and extracts clean TSV data matrices.

# 1. Normalize indels (left-align and split multi-allelic sites)
bcftools norm -m-both -f GRCh38.fa raw_calls.vcf.gz -Oz -o norm.vcf.gz
bcftools index norm.vcf.gz

# 2. Filter for high-confidence PASS variants (Depth >= 20, Genotype Quality >= 30)
bcftools filter -i 'FILTER=="PASS" && QUAL>=30 && FORMAT/DP>=20 && FORMAT/GQ>=30' \
  norm.vcf.gz -Oz -o filtered_highconf.vcf.gz
bcftools index filtered_highconf.vcf.gz

# 3. Query specific variant coordinates and genotypes into flat tabular format
bcftools query -f '%CHROM\t%POS\t%REF\t%ALT[\t%SAMPLE=%GT:%DP]\n' \
  filtered_highconf.vcf.gz > genotype_matrix.tsv