wip

README.md

@@ -193,28 +193,33 @@ The following `FORMAT` fields are written:
 Overall, the command will emit the following header rows in addition to the `##contig=<ID=.,length=.>` lines.
 
 ```
-##fileformat=VCFv4.2
-##FILTER=<ID=PASS,Description="All filters passed">
+##fileformat=VCFv4.4
 ##INFO=<ID=IMPRECISE,Number=0,Type=Flag,Description="Imprecise structural variation">
-##INFO=<ID=chr2,Number=1,Type=String,Description="Second chromosome, if not equal to CHROM">
 ##INFO=<ID=END,Number=1,Type=Integer,Description="End position of the longest variant described in this record">
 ##INFO=<ID=SVTYPE,Number=1,Type=String,Description="Type of structural variant">
 ##INFO=<ID=SVLEN,Number=A,Type=Integer,Description="Length of structural variant">
-##INFO=<ID=SVCLAIM,Number=A,Type=String,Description="Claim made by the structural variant call. Valid values are D, J, DJ for abunda
-##INFO=<ID=annsv,Number=.,Type=String,Description="Effect annotations: 'Allele | Annotation | Gene_Name | Gene_ID'">
+##INFO=<ID=SVCLAIM,Number=A,Type=String,Description="Claim made by the structural variant call. Valid values are D, J, DJ for abundance, adjacency and both respectively">
+##INFO=<ID=chr2,Number=1,Type=String,Description="Second chromosome, if not equal to CHROM">
+##INFO=<ID=annsv,Number=1,Type=String,Description="Effect annotations: 'Allele | Annotation | Gene_Name | Gene_ID'">
+##FILTER=<ID=PASS,Description="All filters passed">
+##FORMAT=<ID=GQ,Number=1,Type=Integer,Description="Conditional genotype quality">
 ##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">
-##FORMAT=<ID=GQ,Number=1,Type=Integer,Description="Genotype Quality">
 ##FORMAT=<ID=pec,Number=1,Type=Integer,Description="Total coverage with paired-end reads">
-##FORMAT=<ID=pev,Number=1,Type=Integer,Description="Variant paired-end read support">
+##FORMAT=<ID=pev,Number=1,Type=Integer,Description="Paired-end reads supporting the variant">
 ##FORMAT=<ID=src,Number=1,Type=Integer,Description="Total coverage with split reads">
-##FORMAT=<ID=srv,Number=1,Type=Integer,Description="Variant split reads support">
-##FORMAT=<ID=amq,Number=1,Type=Integer,Description="Average mapping quality over SV">
+##FORMAT=<ID=srv,Number=1,Type=Integer,Description="Split reads supporting the variant">
+##FORMAT=<ID=amq,Number=1,Type=Float,Description="Average mapping quality over the variant">
 ##FORMAT=<ID=cn,Number=1,Type=Integer,Description="Copy number of the variant in the sample">
-##FORMAT=<ID=anc,Number=1,Type=Float,Description="Average normalized coverage of the variant in the sample">
-##FORMAT=<ID=pc,Number=1,Type=Integer,Description="Point count for CNV call (windows/targets/probes)">
-##x-varfish-version=<ID=varfish-server-worker,Version=x.y.z>
-##x-varfish-version=<ID=orig-caller,Name=DragenSV,Version=07.021.624.3.10.4>
-##x-varfish-version=<ID=orig-caller,Name=Delly2,Version=1.1.7>
+##FORMAT=<ID=anc,Number=1,Type=Float,Description="Average normalized coverage over the variant in the sample">
+##FORMAT=<ID=pc,Number=1,Type=Integer,Description="Point count (windows/targets/probes)">
+##SAMPLE=<ID=index,Sex="Male",Disease="Affected">
+##SAMPLE=<ID=father,Sex="Male",Disease="Unaffected">
+##SAMPLE=<ID=mother,Sex="Female",Disease="Unaffected">
+##PEDIGREE=<ID=index,Father="father",Mother="mother">
+##PEDIGREE=<ID=father>
+##PEDIGREE=<ID=mother>
+##x-varfish-version=<ID=varfish-server-worker,Version="x.y.z">
+##x-varfish-version=<ID=Delly,Name="Delly",Version="1.1.3">
 ```
 
 # Developer Information

src/common.rs

@@ -8,12 +8,12 @@ use std::{
 };
 
 use byte_unit::Byte;
-use clap_verbosity_flag::{InfoLevel, Verbosity};
-
 use clap::Parser;
+use clap_verbosity_flag::{InfoLevel, Verbosity};
 use flate2::{bufread::MultiGzDecoder, write::GzEncoder, Compression};
 use hgvs::static_data::Assembly;
 use indexmap::IndexMap;
+use noodles_vcf as vcf;
 
 /// Commonly used command line arguments.
 #[derive(Parser, Debug)]
@@ -386,3 +386,138 @@ mod test {
         Ok(())
     }
 }
+
+/// Return the version of the `varfish-server-worker` crate and `x.y.z` in tests.
+pub fn worker_version() -> &'static str {
+    if cfg!(test) {
+        "x.y.z"
+    } else {
+        env!("CARGO_PKG_VERSION")
+    }
+}
+
+/// Add contigs for GRCh37.
+pub fn add_contigs_37(
+    builder: vcf::header::Builder,
+) -> Result<vcf::header::Builder, anyhow::Error> {
+    use vcf::header::record::value::map::Contig;
+    use vcf::header::record::value::Map;
+
+    let mut builder = builder;
+
+    let specs: &[(&str, usize); 25] = &[
+        ("1", 249250621),
+        ("2", 243199373),
+        ("3", 198022430),
+        ("4", 191154276),
+        ("5", 180915260),
+        ("6", 171115067),
+        ("7", 159138663),
+        ("8", 146364022),
+        ("9", 141213431),
+        ("10", 135534747),
+        ("11", 135006516),
+        ("12", 133851895),
+        ("13", 115169878),
+        ("14", 107349540),
+        ("15", 102531392),
+        ("16", 90354753),
+        ("17", 81195210),
+        ("18", 78077248),
+        ("19", 59128983),
+        ("20", 63025520),
+        ("21", 48129895),
+        ("22", 51304566),
+        ("X", 155270560),
+        ("Y", 59373566),
+        ("MT", 16569),
+    ];
+
+    for (contig, length) in specs {
+        builder = builder.add_contig(
+            contig
+                .parse()
+                .map_err(|_| anyhow::anyhow!("invalid contig: {}", contig))?,
+            Map::<Contig>::builder()
+                .set_length(*length)
+                .insert(
+                    "assembly"
+                        .parse()
+                        .map_err(|_| anyhow::anyhow!("invalid key: assembly"))?,
+                    "GRCh37",
+                )
+                .insert(
+                    "species"
+                        .parse()
+                        .map_err(|_| anyhow::anyhow!("invalid key: species"))?,
+                    "Homo sapiens",
+                )
+                .build()?,
+        );
+    }
+
+    Ok(builder)
+}
+
+/// Add contigs for GRCh38.
+pub fn add_contigs_38(
+    builder: vcf::header::Builder,
+) -> Result<vcf::header::Builder, anyhow::Error> {
+    use vcf::header::record::value::map::Contig;
+    use vcf::header::record::value::Map;
+
+    let mut builder = builder;
+
+    let specs: &[(&str, usize); 25] = &[
+        ("chr1", 248956422),
+        ("chr2", 242193529),
+        ("chr3", 198295559),
+        ("chr4", 190214555),
+        ("chr5", 181538259),
+        ("chr6", 170805979),
+        ("chr7", 159345973),
+        ("chr8", 145138636),
+        ("chr9", 138394717),
+        ("chr10", 133797422),
+        ("chr11", 135086622),
+        ("chr12", 133275309),
+        ("chr13", 114364328),
+        ("chr14", 107043718),
+        ("chr15", 101991189),
+        ("chr16", 90338345),
+        ("chr17", 83257441),
+        ("chr18", 80373285),
+        ("chr19", 58617616),
+        ("chr20", 64444167),
+        ("chr21", 46709983),
+        ("chr22", 50818468),
+        ("chrX", 156040895),
+        ("chrY", 57227415),
+        ("chrM", 16569),
+    ];
+
+    for (contig, length) in specs {
+        builder = builder.add_contig(
+            contig
+                .parse()
+                .map_err(|_| anyhow::anyhow!("invalid contig: {}", contig))?,
+            Map::<Contig>::builder()
+                .set_length(*length)
+                .insert(
+                    "assembly"
+                        .parse()
+                        .map_err(|_| anyhow::anyhow!("invalid key: assembly"))?,
+                    "GRCh38",
+                )
+                .insert(
+                    "species"
+                        .parse()
+                        .map_err(|_| anyhow::anyhow!("invalid key: species"))?,
+                    "Homo sapiens",
+                )
+                .build()?,
+        );
+    }
+
+    Ok(builder)
+}

src/db/mk_inhouse/cli.rs

@@ -20,7 +20,7 @@ use crate::{
         build_chrom_map, open_read_maybe_gz, open_write_maybe_gz, read_lines, trace_rss_now,
         GenomeRelease, CHROMS,
     },
-    sv::query::schema::SvType,
+    strucvars::query::schema::SvType,
 };
 
 /// Create one file with records for each chromosome and SV type.

src/db/mk_inhouse/input.rs

@@ -2,7 +2,7 @@
 
 use serde::{de::IntoDeserializer, Deserialize, Deserializer, Serialize};
 
-use crate::sv::query::schema::{StrandOrientation, SvType};
+use crate::strucvars::query::schema::{StrandOrientation, SvType};
 
 /// Representation of the fields from the `StructuralVariant` table from
 /// VarFish Server that we need for building the background records.

src/db/mk_inhouse/output.rs

@@ -2,7 +2,7 @@
 
 use serde::{Deserialize, Serialize};
 
-use crate::sv::query::schema::{StrandOrientation, SvType};
+use crate::strucvars::query::schema::{StrandOrientation, SvType};
 
 use super::input::Record as InputRecord;
 

src/db/to_bin/clinvar/input.rs

@@ -3,7 +3,7 @@
 //! Note that not the full model is implemented, only the parts that are needed for the
 //! conversion of the ClinVar structural variants.
 
-use crate::sv::query::clinvar::pbs::{Pathogenicity, VariationType};
+use crate::strucvars::query::clinvar::pbs::{Pathogenicity, VariationType};
 
 /// Accession of a ClinVar record.
 #[derive(Debug, serde::Deserialize, serde::Serialize)]

src/db/to_bin/clinvar/mod.rs

@@ -7,7 +7,7 @@ use thousands::Separable;
 
 use crate::{
     common::{build_chrom_map, open_read_maybe_gz, trace_rss_now},
-    sv::query::clinvar::pbs::{Pathogenicity, SvDatabase, SvRecord},
+    strucvars::query::clinvar::pbs::{Pathogenicity, SvDatabase, SvRecord},
 };
 
 pub mod input;
@@ -71,7 +71,7 @@ fn convert_jsonl_to_protobuf(
         for measure in &record.reference_clinvar_assertion.measures.measures {
             // convert from JSONL to protocolbuffers: variation type
             let variation_type: Result<
-                crate::sv::query::clinvar::pbs::VariationType,
+                crate::strucvars::query::clinvar::pbs::VariationType,
                 anyhow::Error,
             > = measure.r#type.try_into();
             let variation_type = if let Ok(variation_type) = variation_type {

src/db/to_bin/vardbs/input.rs

@@ -5,7 +5,7 @@ use serde::Deserialize;
 use tracing::error;
 
 use crate::db::mk_inhouse::output::Record as InhouseDbRecord;
-use crate::sv::query::schema::SvType;
+use crate::strucvars::query::schema::SvType;
 
 /// dbVar database record as read from TSV file.
 #[derive(Debug, Deserialize)]

src/db/to_bin/vardbs/mod.rs

@@ -13,7 +13,7 @@ use crate::common::{build_chrom_map, open_read_maybe_gz, trace_rss_now};
 use crate::db;
 use crate::db::mk_inhouse::output::Record as InhouseDbRecord;
 use crate::db::pbs::{BackgroundDatabase, BgDbRecord};
-use crate::sv::query::schema::SvType;
+use crate::strucvars::query::schema::SvType;
 
 use self::input::InputRecord;
 

src/main.rs

@@ -3,7 +3,7 @@
 pub mod common;
 pub mod db;
 pub mod seqvars;
-pub mod sv;
+pub mod strucvars;
 
 use clap::{Args, Parser, Subcommand};
 use console::{Emoji, Term};
@@ -32,8 +32,8 @@ struct Cli {
 enum Commands {
     /// Database building related commands.
     Db(Db),
-    /// SV filtration related commands.
-    Sv(Sv),
+    /// Structural variant related commands.
+    Strucvars(Strucvars),
     /// Sequence variant related commands.
     Seqvars(Seqvars),
 }
@@ -58,16 +58,17 @@ enum DbCommands {
 /// Parsing of "sv *" sub commands.
 #[derive(Debug, Args)]
 #[command(args_conflicts_with_subcommands = true)]
-struct Sv {
+struct Strucvars {
     /// The sub command to run
     #[command(subcommand)]
-    command: SvCommands,
+    command: StrucvarsCommands,
 }
 
 /// Enum supporting the parsing of "sv *" sub commands.
 #[derive(Debug, Subcommand)]
-enum SvCommands {
-    Query(sv::query::Args),
+enum StrucvarsCommands {
+    Ingest(strucvars::ingest::Args),
+    Query(strucvars::query::Args),
 }
 
 /// Parsing of "seqvars *" sub commands.
@@ -121,9 +122,12 @@ fn main() -> Result<(), anyhow::Error> {
                     seqvars::ingest::run(&cli.common, args)?;
                 }
             },
-            Commands::Sv(sv) => match &sv.command {
-                SvCommands::Query(args) => {
-                    sv::query::run(&cli.common, args)?;
+            Commands::Strucvars(strucvars) => match &strucvars.command {
+                StrucvarsCommands::Ingest(args) => {
+                    strucvars::ingest::run(&cli.common, args)?;
+                }
+                StrucvarsCommands::Query(args) => {
+                    strucvars::query::run(&cli.common, args)?;
                 }
             },
         }

src/seqvars/ingest/mod.rs

@@ -2,7 +2,7 @@
 
 use std::sync::{Arc, OnceLock};
 
-use crate::common::{self, open_read_maybe_gz, open_write_maybe_gz, GenomeRelease};
+use crate::common::{self, open_read_maybe_gz, open_write_maybe_gz, worker_version, GenomeRelease};
 use mehari::annotate::seqvars::provider::MehariProvider;
 use noodles_vcf as vcf;
 use thousands::Separable;
@@ -33,15 +33,6 @@ pub struct Args {
     pub path_out: String,
 }
 
-/// Return the version of the `varfish-server-worker` crate and `x.y.z` in tests.
-fn worker_version() -> &'static str {
-    if cfg!(test) {
-        "x.y.z"
-    } else {
-        env!("CARGO_PKG_VERSION")
-    }
-}
-
 /// Return path component fo rth egiven assembly.
 pub fn path_component(genomebuild: GenomeRelease) -> &'static str {
     match genomebuild {