config.yaml

# config.yaml

####################################################Prepare these paths before running the install_utils.sh script####################################################
tmp_dir: "/paedyl01/disk1/yangyxt/test_tmp/test_acmg"
assembly: hg19   # Change this you need to re-run the install_utils.sh script
ref_genome: "/paedyl01/disk1/yangyxt/indexed_genome/ucsc.hg19.fasta"  # Be consistent of the assembly tag
conda_env_yaml: "acmg_conda_rough.yml"  # Included in the GitHub repo, not recommend to change
conda_env_name: "priva_acmg"   # If choose alternative env name, pls update the yaml file accordingly that will be used by the install_utils.sh script
hg38_hg19_chain: "hg38ToHg19.over.chain.gz"   # Included in the GitHub repo, not recommend to change

# Below specify the major directories to store the annotation resources
gnomad_vcf_dir: "/paedyl01/disk1/yangyxt/public_data/gnomAD/v4_dataset"  # The directory to store the gnomAD related annotation resources
clinvar_vcf_dir: "/paedyl01/disk1/yangyxt/public_data/clinvar"  # The directory to store the ClinVar related annotation resources
vep_installed_version: "113"
vep_cache_dir: "/paedyl01/disk1/yangyxt/public_data/VEP_caches"  # the cache directory for VEP
vep_plugins_dir: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins"  # the directory containing the VEP plugins executable files
vep_plugins_cachedir: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches"   # the directory for storing the VEP plugin caches
cadd_parent_dir: "/paedyl01/disk1/yangyxt/Tools/CADD"  # The parent directory to store the CADD scripts and cache files
####################################################Ready to RUN install_utils.sh script####################################################


####################################################Prepare these paths before running PriVA workflow####################################################
input_vcf: "/paedyl01/disk1/yangyxt/test_acmg_auto/TEST_FAM.filtered.vcf.gz"  # Below we specify the input vcf file in absolute paths
ped_file: "/paedyl01/disk1/yangyxt/test_acmg_auto/TEST_FAM.ped"  # The PED file containing the identical sample set included in the input VCF file
output_dir: "/paedyl01/disk1/yangyxt/test_acmg_auto"  # Output directory in absolute paths, per-family outputs will be stored directly in it

hub_file_dir: "/paedyl01/disk1/yangyxt/ngs_summary"   # the directory for storing the hub files for annotation caching
hub_vcf_file: "hg19.anno.vcf.gz"    # hub file name for annotation caching
hub_cadd_file: "hg19.cadd.tsv.gz"    # hub file name for CADD annotation caching

extreme_rare_PAF: 0.0001
exp_disease_incidence: 0.001
threads: 20          # Total threads for annotation / non-family steps
threads_per_fam: 4   # Threads for EACH family filtering/prioritization job
af_cutoff: 0.05      # Allele frequency cutoff for common variants

alt_disease_vcf: "/paedyl01/disk1/yangyxt/public_data/ShortStatue/ShortStature.hg19.vcf.gz"  # abs path to the VCF file containing the variants from alternative disease patients, make sure it is left-normalized and biallelic, optional
control_vcf: ""   # VCF file containing the variants from control individuals, make sure it is left-normalized and biallelic, optional
pp1_vcf: ""      # Cohort VCF file containing patients and their family members for PP1 (cosegregation), optional
pp1_ped: ""      # PED file for the PP1 cohort VCF file, optional
####################################################You can adjust above argument values for each cohort####################################################


####################################################Below paths will be determined automatically by install_utils.sh####################################################
# Below we specify the path to the gnomAD vcf file for chromosome X and we'll use that template to deduce the paths to other chromosome VCF files, the paths should be relatively constant
gnomad_vcf_chrX: "/paedyl01/disk1/yangyxt/public_data/gnomAD/v4_dataset/gnomad.joint.v4.1.sites.hg19.chrX.vcf.gz"

# Below we specify the path to the ClinVar VCF file, the path should be relatively constant
clinvar_vcf: "/paedyl01/disk1/yangyxt/public_data/clinvar/clinvar.hg19.vep.vcf.gz"
clinvar_aa_stat: "/paedyl01/disk1/yangyxt/public_data/clinvar/clinvar.hg19.vep.aa_change.stats.pkl"
clinvar_splice_stat: "/paedyl01/disk1/yangyxt/public_data/clinvar/clinvar.hg19.vep.splice_change.stats.pkl"
clinvar_patho_af_stat: "/paedyl01/disk1/yangyxt/public_data/clinvar/clinvar.hg19.vep.patho_af_stat.pkl"
clinvar_patho_exon_af_stat: "/paedyl01/disk1/yangyxt/public_data/clinvar/clinvar.hg19.vep.patho_exon_af_stat.pkl"
clinvar_gene_stat: "clinvar_2star_stats.pkl.gz"

# The UTRAnnotator plugin cache file
utr_annotator_file: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/UTRannotator/uORF_5UTR_GRCh37_PUBLIC.txt"
# The LOEUF plugin cache file
loeuf_prescore: "loeuf_dataset.tsv.gz"

# The LOFTEE plugin cache file and LOFTEE repository
loftee_parent_dir: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins/LoFtee_repos"
loftee_repo: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins/LoFtee_repos/loftee-hg19"
human_ancestor_fasta: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/LoFtee/hg19/human_ancestor.fa.gz"
loftee_conservation_file: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/LoFtee/hg19/phylocsf_gerp.sql"  # The GERP SQL file must not be compressed.
gerp_bigwig: "" # Below is only available for GRCh38

# The AlphaMissense plugin cache file
alphamissense_prescore: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/AlphaMissense_hg19.tsv.gz
alphamissense_vcf: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/AlphaMissense_hg19.vcf.gz
alphamissense_vep_vcf: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/AlphaMissense_hg19.vep.vcf.gz
alphamissense_pd_stat: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/AlphaMissense_hg19.vep.prot.domain.stats.pkl
alphamissense_tranx_domain_map: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/transcript_pp_domain_mapping.hg19.pkl
alphamissense_intolerant_domains: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/domain_tolerance_analysis.hg19.tsv
alphamissense_intolerant_motifs: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/AlphaMissense_hg19.kde.pkl

# Final intolerant domains combining results from AlphaMissense and ClinVar
all_intolerant_domains: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/all_intolerant_domains.hg19.pkl

# Combined PM1 regions (DAS + HCSeeker + RMC)
pm1_regions_pkl: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/AlphaMissense/pm1_regions.hg19.pkl

# The Conservation cache file
conservation_file: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/Conservation/gerp_conservation_scores.homo_sapiens.GRCh37.bw

# The SpliceVault plugin cache file
splicevault_url: "https://ftp.ensembl.org/pub/current_variation/SpliceVault/SpliceVault_data_GRCh38.tsv.gz"
splicevault_prescore: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/SpliceVault/SpliceVault_data_GRCh37.tsv.gz"

# CADD directory storing its executable and cache files
cadd_base_dir: /paedyl01/disk1/yangyxt/Tools/CADD/CADD-scripts-1.7.2
####################################################Above paths will be determined automatically by install_utils.sh####################################################


####################################################Below will be determined manually by your reply by install_utils.sh####################################################
# These cache files are too large you need to register a free account in Illumina to download them upon prompt of the install_utils.sh execution(size up to several hundreds of GB)

# The PrimateAI plugin prescore file
primateai_prescore: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/PrimateAI/hg19/PrimateAI_scores_v0.2_GRCh37_sorted.tsv.bgz

# The SpliceAI plugin cache file, snv and indel
spliceai_snv_prescore: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/SpliceAI/hg19/spliceai_scores.raw.snv.hg19.vcf.gz
spliceai_indel_prescore: /paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/SpliceAI/hg19/spliceai_scores.raw.indel.hg19.vcf.gz
####################################################Above will be determined manually by your reply by install_utils.sh####################################################



####################################################Below are constant values, please report issue when one of them fails####################################################
repeat_region_file_name: "ucsc.hg19.fasta.repeat.bed.gz"  # Included in the GitHub repo, can be customized if needed

# Below we specify the path to the CADD base directory
cadd_zip_download_url: "https://github.com/kircherlab/CADD-scripts/archive/refs/tags/v1.7.2.zip"
cadd_GRCh37_anno_md5: "22549309c65de8d25e903695aa4f749f"
cadd_GRCh38_anno_md5: "205d3e702df3565efb424e2ca80c9d25"
cadd_GRCh37_snv_anno_url: "https://krishna.gs.washington.edu/download/CADD/v1.7/GRCh37/whole_genome_SNVs_inclAnno.tsv.gz"
cadd_GRCh37_snv_anno_md5: "1b89cfd7da4e2a8f336c1ec33a6f44eb"
cadd_GRCh37_indel_anno_url: "https://krishna.gs.washington.edu/download/CADD/v1.7/GRCh37/gnomad.genomes-exomes.r4.0.indel_inclAnno.tsv.gz"
cadd_GRCh37_indel_anno_md5: "3d88f8fde9df07060b4f8fe07a0d5d5d"
cadd_GRCh38_snv_anno_url: "https://krishna.gs.washington.edu/download/CADD/v1.7/GRCh38/whole_genome_SNVs_inclAnno.tsv.gz"
cadd_GRCh38_snv_anno_md5: "ed9d3e029686375e2ac80e96b5d0067a"
cadd_GRCh38_indel_anno_url: "https://krishna.gs.washington.edu/download/CADD/v1.7/GRCh38/gnomad.genomes.r4.0.indel_inclAnno.tsv.gz"
cadd_GRCh38_indel_anno_md5: "00ce2d2431a6ec3324b23ff625f4e0cd"
cadd_version: "v1.7"

# InterPro paths
interpro_xml_gz_url: "https://ftp.ebi.ac.uk/pub/databases/interpro/current_release/interpro.xml.gz"
interpro_mapping_pickle: "Interpro_entry_mapping.pkl.gz"

# ClinGen paths
clingen_map: "/paedyl01/disk1/yangyxt/PriVA/data/clingen/clingen_map.hg19.pkl.gz"
gene_dosage_sensitivity: "/paedyl01/disk1/yangyxt/PriVA/data/clingen/gene_dosage_sensitivity.hg19.tsv"

# For any variant ClinGen has curated, ClinGen's applied evidence codes replace
# the criteria PriVA derived. This is the production setting: an expert panel's
# verdict outranks PriVA's own reasoning about the same variant.
#
# Set this to false ONLY when benchmarking PriVA against ClinGen-curated
# variants. With the override on, that comparison is circular -- PriVA would be
# scored on how well it reproduces answers it was handed, and would agree with
# ClinGen almost perfectly on exactly the variants used to measure it.
#
# Anything other than the exact string false leaves the override on, so a config
# that predates this key keeps production behaviour.
apply_clingen_override: true

# Ensembl CDS FASTA for alternative start codon detection (used for start_lost variants)
# Download URLs (always latest via "current" symlinks, installed by install_utils.sh):
#   GRCh38: https://ftp.ensembl.org/pub/current_fasta/homo_sapiens/cds/Homo_sapiens.GRCh38.cds.all.fa.gz
#   GRCh37: https://ftp.ensembl.org/pub/grch37/current/fasta/homo_sapiens/cds/Homo_sapiens.GRCh37.cds.all.fa.gz
cds_fasta_file: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/Ensembl_CDS/Homo_sapiens.GRCh37.cds.all.fa.gz"

# MaveDB paths
mavedb_url: "https://ftp.ensembl.org/pub/current_variation/MaveDB/MaveDB_variants.tsv.gz"
mavedb_file: "/paedyl01/disk1/yangyxt/public_data/VEP_plugins_caches/MaveDB/MaveDB_variants.hg38.tsv.gz"

# Gene mechanism / DDG2P-G2P cache paths
# Built by: bash scripts/install_utils.sh gene_pathogenic_mechanism_cache_install config.yaml
# One flat folder. Only canonical runtime caches and required cache-build inputs
# are kept in the repository; audit byproducts land beside them and are ignored.
# The raw download tree is shared outside the repository, so it is named
# absolutely.
gene_mechanism_cache_dir: "/paedyl01/disk1/yangyxt/PriVA/data/patho_mechanism"
gene_mechanism_raw_dir: "/paedyl01/disk1/yangyxt/public_data/gene_pathogenic_mechanism/raw"
gene_mechanism_builder_script: "/paedyl01/disk1/yangyxt/PriVA/scripts/build_gene_pathogenic_mechanism_cache.py"
gene_nonlof_mechanism_builder_script: /paedyl01/disk1/yangyxt/PriVA/scripts/build_gene_nonlof_mechanism_cache.py
gene_nonlof_mechanism_schema: /paedyl01/disk1/yangyxt/PriVA/data/patho_mechanism/gene_nonlof_mechanism_curated_assertions.schema.json
# Step 3 reads these two runtime caches directly. The condition-mechanism TSV
# below is only a build input to the HPO-framed cache.
gene_nonlof_mechanism_json: /paedyl01/disk1/yangyxt/PriVA/data/patho_mechanism/gene_nonlof_mechanism_curated_assertions.json.gz
hpo_condition_mechanism_builder_script: /paedyl01/disk1/yangyxt/PriVA/scripts/build_hpo_condition_mechanism_cache.py
hpo_condition_mechanism_json: /paedyl01/disk1/yangyxt/PriVA/data/patho_mechanism/hpo_condition_mechanism_cache.json.gz
ddg2p_mechanism_evidence: /paedyl01/disk1/yangyxt/PriVA/data/patho_mechanism/gene_pathogenic_mechanism_evidence.tsv
hgnc_table: /paedyl01/disk1/yangyxt/PriVA/data/hgnc/non_alt_loci_set.tsv
g2p_ddg2p_url: "https://www.ebi.ac.uk/gene2phenotype/api/panel/all/download"
orphadata_gene_disease_url: "https://www.orphadata.com/data/xml/en_product6.xml"
clingen_dosage_grch38_url: "https://ftp.clinicalgenome.org/ClinGen_gene_curation_list_GRCh38.tsv"
panelapp_panels_url: "https://panelapp.genomicsengland.co.uk/api/v1/panels/"

# GoFCards exact variant-level GOF cache paths
# One deployed cache, built in two steps. The first normalizes the GoFCards
# catalogue into <gofcards_workdir>/gofcards_exact_gof.normalized.json.gz, a
# build intermediate; the second nests each variant's ClinVar conditions into it
# and writes the file below, which is what every consumer reads.
# mechanism_resource_install runs both, and runs a third step between them that
# downloads the weekly ClinVar VCV XML the second step needs. To run them by
# hand, keep this order:
#   bash scripts/install_utils.sh gofcards_exact_gof_cache_install     config.yaml
#   bash scripts/install_utils.sh gene_nonlof_mechanism_cache_install  config.yaml sources-only
#   bash scripts/install_utils.sh gofcards_clinvar_injection_install   config.yaml
gofcards_exact_gof_cache: /paedyl01/disk1/yangyxt/PriVA/data/gofcards/gofcards_exact_gof.json.gz
gofcards_public_excel_url: "https://download.genemed.tech/upload/GainFunCards/gofcards_data_download.xlsx"
# Longest the normalized cache may go without being rebuilt to pick up a new
# GoFCards release. It is an upper bound only: the installer also rebuilds as
# soon as the normalization code, the mechanism review table, the HGNC table,
# the chain, either FASTA, or the VEP annotation cache changes.
gofcards_refresh_days: "180"
# How far the deployed cache may trail the weekly ClinVar VCV XML before its
# conditions are injected again. ClinVar ships a release every week and each
# re-injection reads the whole 5.8 GB file, so chasing every release costs hours
# for almost no change. Three months.
gofcards_clinvar_reinjection_lag_days: "90"
gofcards_workdir: "/paedyl01/disk1/yangyxt/PriVA/data/gofcards/build_work"
gofcards_hg38_fasta: "/paedyl01/disk1/yangyxt/indexed_genome/hg38/ucsc.hg38.fasta"
gofcards_hg19_to_hg38_chain: "/paedyl01/disk1/yangyxt/public_data/liftOver/hg19ToHg38.over.chain"
gofcards_clinvar_injector_script: "/paedyl01/disk1/yangyxt/PriVA/scripts/inject_clinvar_into_gofcards.py"
clinvar_vcv_xml: "/paedyl01/disk1/yangyxt/public_data/gene_pathogenic_mechanism/raw/clinvar_vcv/ClinVarVCVRelease_00-latest_weekly.xml.gz"
gofcards_builder_script: "/paedyl01/disk1/yangyxt/PriVA/scripts/build_gofcards_exact_gof_cache.py"
gofcards_mechanism_review_tsv: "/paedyl01/disk1/yangyxt/PriVA/data/gofcards/gofcards_mechanism_reviews.tsv"

# HPO/MONDO disease-scope resources
# Built by: bash scripts/install_utils.sh mondo_hpo_scope_install config.yaml
# Releases are pinned so a normal reinstall is reproducible. Update the release,
# URL, and checksum together when intentionally refreshing MONDO.
hpo_release: "v2026-06-23"
hpo_raw_dir: "/paedyl01/disk1/yangyxt/PriVA/data/hpo/raw/v2026-06-23"
hpo_genes_to_phenotype_url: "https://github.com/obophenotype/human-phenotype-ontology/releases/download/v2026-06-23/genes_to_phenotype.txt"
hpo_phenotype_hpoa_url: "https://github.com/obophenotype/human-phenotype-ontology/releases/download/v2026-06-23/phenotype.hpoa"
hpo_assertion_builder_script: "/paedyl01/disk1/yangyxt/PriVA/scripts/collapse_HPO_anno.py"
hpo_assertions: /paedyl01/disk1/yangyxt/PriVA/data/hpo/genes_to_phenotype.assertions.tsv.gz
mondo_release: "v2026-07-06"
mondo_url: "https://github.com/monarch-initiative/mondo/releases/download/v2026-07-06/mondo-simple.obo"
mondo_sha256: "7e28f3fba97b2f474b70dba54d9b719d07c3d84e5e682189146a05dbedf755aa"
mondo_obo: /paedyl01/disk1/yangyxt/PriVA/data/mondo/mondo-simple.v2026-07-06.obo
mondo_scope_builder_script: "/paedyl01/disk1/yangyxt/PriVA/scripts/build_mondo_disease_scope.py"
mondo_disease_scope_registry: /paedyl01/disk1/yangyxt/PriVA/data/mondo/disease_scope.tsv.gz
mondo_disease_scope_overrides: "/paedyl01/disk1/yangyxt/PriVA/data/mondo/disease_scope_overrides.tsv"


####################################################pext (proportion expressed across transcripts) Configuration####################################################
# pext from gnomAD, based on GTEx v10 expression data (updated Nov 2024)
# Reference: Cummings et al., "Transcript expression-aware annotation improves rare variant interpretation", Nature 2020
# If pext_enabled is true, the mean pext file is always downloaded and annotated via VEP --custom bigwig
# Optionally specify comma-delimited tissue names for tissue-specific pext annotation
#
# NOTE: Tissue names differ slightly between hg19 and hg38 assemblies!
#
# COMMON TISSUES (both assemblies):
#   Adipose_Subcutaneous, AdrenalGland, Artery_Aorta, Artery_Coronary, Artery_Tibial,
#   Bladder, Brain_Amygdala, Brain_CerebellarHemisphere, Brain_Cerebellum, Brain_Cortex,
#   Brain_Hippocampus, Brain_Hypothalamus, Brain_Substantianigra, Breast_MammaryTissue,
#   Cells_EBV_transformedlymphocytes, Colon_Sigmoid, Colon_Transverse,
#   Esophagus_GastroesophagealJunction, Esophagus_Mucosa, Esophagus_Muscularis,
#   Heart_AtrialAppendage, Heart_LeftVentricle, Kidney_Cortex, Liver, Lung,
#   MinorSalivaryGland, Muscle_Skeletal, Nerve_Tibial, Ovary, Pancreas, Pituitary,
#   Prostate, SmallIntestine_TerminalIleum, Spleen, Stomach, Testis, Thyroid,
#   Uterus, Vagina, WholeBlood
#
# hg38/GRCh38 ADDITIONAL:
#   Adipose_Visceral_Omentum, Brain_Anteriorcingulatecortex_BA24, Brain_Caudate_basalganglia,
#   Brain_FrontalCortex_BA9, Brain_Nucleusaccumbens_basalganglia, Brain_Putamen_basalganglia,
#   Brain_Spinalcord_cervicalc_1, Cells_Culturedfibroblasts, Skin_NotSunExposed_Suprapubic,
#   Skin_SunExposed_Lowerleg
#
# hg19/GRCh37 ADDITIONAL (note trailing underscores on some):
#   Adipose_Visceral_Omentum_, Brain_Anteriorcingulatecortex_BA24_, Brain_Caudate_basalganglia_,
#   Brain_FrontalCortex_BA9_, Brain_Nucleusaccumbens_basalganglia_, Brain_Putamen_basalganglia_,
#   Brain_Spinalcord_cervicalc_1_, Cells_Transformedfibroblasts, Cervix_Ectocervix,
#   Cervix_Endocervix, FallopianTube, Skin_NotSunExposed_Suprapubic_, Skin_SunExposed_Lowerleg_
#
pext_enabled: true                                           # Set to false to skip pext annotation entirely
pext_tissues: "Spleen,WholeBlood,Cells_EBV_transformedlymphocytes"  # Immune-related tissues for hg19
# pext_tissues: ""                                           # Leave empty to use only mean pext (recommended for most use cases)

# Auto-populated by install_utils.sh:
pext_mean_bw: ""                                             # Path to mean pext BigWig file
pext_tissue_bw_dir: ""                                       # Directory containing tissue-specific BigWig files