Skip to content

Commit

Permalink
fix(derive/endedness): add logic for 0x1 bit
Browse files Browse the repository at this point in the history
  • Loading branch information
@a-frantz
a-frantz committed Feb 8, 2024
1 parent 89b60e0 commit d5b5b5c
Show file tree
Hide file tree
Showing 2 changed files with 71 additions and 3 deletions.
23 changes: 22 additions & 1 deletion src/derive/command/endedness.rs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -122,7 +122,24 @@ pub fn derive(args: DeriveEndednessArgs) -> anyhow::Result<()> {

let overall_rg = Arc::clone(&OVERALL);

if record.flags().is_first_segment() && !record.flags().is_last_segment() {
if !record.flags().is_segmented() {
ordering_flags.entry(overall_rg).and_modify(|e| {
e.unsegmented += 1;
});

ordering_flags
.entry(read_group)
.and_modify(|e| {
e.unsegmented += 1;
})
.or_insert(OrderingFlagsCounts {
unsegmented: 1,
first: 0,
last: 0,
both: 0,
neither: 0,
});
} else if record.flags().is_first_segment() && !record.flags().is_last_segment() {
ordering_flags.entry(overall_rg).and_modify(|e| {
e.first += 1;
});
Expand All @@ -133,6 +150,7 @@ pub fn derive(args: DeriveEndednessArgs) -> anyhow::Result<()> {
e.first += 1;
})
.or_insert(OrderingFlagsCounts {
unsegmented: 0,
first: 1,
last: 0,
both: 0,
Expand All @@ -149,6 +167,7 @@ pub fn derive(args: DeriveEndednessArgs) -> anyhow::Result<()> {
e.last += 1;
})
.or_insert(OrderingFlagsCounts {
unsegmented: 0,
first: 0,
last: 1,
both: 0,
Expand All @@ -165,6 +184,7 @@ pub fn derive(args: DeriveEndednessArgs) -> anyhow::Result<()> {
e.both += 1;
})
.or_insert(OrderingFlagsCounts {
unsegmented: 0,
first: 0,
last: 0,
both: 1,
Expand All @@ -181,6 +201,7 @@ pub fn derive(args: DeriveEndednessArgs) -> anyhow::Result<()> {
e.neither += 1;
})
.or_insert(OrderingFlagsCounts {
unsegmented: 0,
first: 0,
last: 0,
both: 0,
Expand Down
51 changes: 49 additions & 2 deletions src/derive/endedness/compute.rs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,9 @@ use crate::utils::read_groups::{OVERALL, UNKNOWN_READ_GROUP};
/// Struct holding the ordering flags for a single read group.
#[derive(Debug, Clone)]
pub struct OrderingFlagsCounts {
/// The number of reads without 0x1 set.
pub unsegmented: usize,

/// The number of reads with the first in template flag set.
pub first: usize,

Expand All @@ -27,6 +30,7 @@ impl OrderingFlagsCounts {
/// Creates a new [`OrderingFlagsCounts`].
pub fn new() -> Self {
OrderingFlagsCounts {
unsegmented: 0,
first: 0,
last: 0,
both: 0,
Expand Down Expand Up @@ -54,6 +58,9 @@ pub struct ReadGroupDerivedEndednessResult {
/// The endedness of this read group or "Unknown".
pub endedness: String,

/// The number of reads without 0x1 set.
pub unsegmented: usize,

/// The f+l- read count.
pub first: usize,

Expand Down Expand Up @@ -84,6 +91,7 @@ impl ReadGroupDerivedEndednessResult {
read_group,
succeeded,
endedness,
unsegmented: counts.unsegmented,
first: counts.first,
last: counts.last,
both: counts.both,
Expand All @@ -103,6 +111,9 @@ pub struct DerivedEndednessResult {
/// The overall endedness of the file or "Unknown".
pub endedness: String,

/// The number of reads without 0x1 set.
pub unsegmented: usize,

/// The overall f+l- read count.
pub first: usize,

Expand Down Expand Up @@ -137,6 +148,7 @@ impl DerivedEndednessResult {
DerivedEndednessResult {
succeeded,
endedness,
unsegmented: counts.unsegmented,
first: counts.first,
last: counts.last,
both: counts.both,
Expand Down Expand Up @@ -234,14 +246,15 @@ fn predict_endedness(
reads_per_template: Option<&f64>,
round_rpt: bool,
) -> ReadGroupDerivedEndednessResult {
let unsegmented = rg_ordering_flags.unsegmented;
let first = rg_ordering_flags.first;
let last = rg_ordering_flags.last;
let both = rg_ordering_flags.both;
let neither = rg_ordering_flags.neither;

// all zeroes (Perform this check before creating the result struct
// so that we don't have to clone the read group name)
if first == 0 && last == 0 && both == 0 && neither == 0 {
if unsegmented == 0 && first == 0 && last == 0 && both == 0 && neither == 0 {
warn!(
"No reads were detected in this read group: {}",
read_group_name
Expand All @@ -263,6 +276,28 @@ fn predict_endedness(
reads_per_template.copied(),
);

// only unsegmented present
if unsegmented > 0 && first == 0 && last == 0 && both == 0 && neither == 0 {
match reads_per_template {
Some(rpt) => {
if *rpt == 1.0 || (round_rpt && rpt.round() as usize == 1) {
result.succeeded = true;
result.endedness = String::from("Single-End");
}
}
None => {
result.succeeded = true;
result.endedness = String::from("Single-End");
}
}
return result;
}
// unsegmented reads are present, and so are other types of reads.
if unsegmented > 0 {
return result;
}
// now unsegmented is guarenteed to be 0

// only first present
if first > 0 && last == 0 && both == 0 && neither == 0 {
return result;
Expand Down Expand Up @@ -347,7 +382,8 @@ pub fn predict(

for (read_group, rg_ordering_flags) in ordering_flags.iter() {
if (*read_group == *UNKNOWN_READ_GROUP)
&& (rg_ordering_flags.first == 0
&& (rg_ordering_flags.unsegmented == 0
&& rg_ordering_flags.first == 0
&& rg_ordering_flags.last == 0
&& rg_ordering_flags.both == 0
&& rg_ordering_flags.neither == 0)
Expand All @@ -363,6 +399,7 @@ pub fn predict(
);
if result.read_group == "overall" {
final_result.endedness = result.endedness;
final_result.unsegmented = result.unsegmented;
final_result.first = result.first;
final_result.last = result.last;
final_result.both = result.both;
Expand All @@ -381,12 +418,14 @@ pub fn predict(
mod tests {
use super::*;

// TODO add tests for unsegmented reads
#[test]
fn test_predict_endedness() {
let mut ordering_flags: HashMap<Arc<String>, OrderingFlagsCounts> = HashMap::new();
ordering_flags.insert(
Arc::clone(&OVERALL),
OrderingFlagsCounts {
unsegmented: 0,
first: 1,
last: 1,
both: 0,
Expand Down Expand Up @@ -435,6 +474,7 @@ mod tests {
ordering_flags.insert(
Arc::clone(&OVERALL),
OrderingFlagsCounts {
unsegmented: 0,
first: 1,
last: 0,
both: 0,
Expand All @@ -458,6 +498,7 @@ mod tests {
ordering_flags.insert(
Arc::clone(&OVERALL),
OrderingFlagsCounts {
unsegmented: 0,
first: 0,
last: 1,
both: 0,
Expand All @@ -481,6 +522,7 @@ mod tests {
ordering_flags.insert(
Arc::clone(&OVERALL),
OrderingFlagsCounts {
unsegmented: 0,
first: 0,
last: 0,
both: 1,
Expand All @@ -504,6 +546,7 @@ mod tests {
ordering_flags.insert(
Arc::clone(&OVERALL),
OrderingFlagsCounts {
unsegmented: 0,
first: 0,
last: 0,
both: 0,
Expand All @@ -527,6 +570,7 @@ mod tests {
ordering_flags.insert(
Arc::clone(&OVERALL),
OrderingFlagsCounts {
unsegmented: 0,
first: 1,
last: 1,
both: 0,
Expand Down Expand Up @@ -589,6 +633,7 @@ mod tests {
ordering_flags.insert(
Arc::clone(&OVERALL),
OrderingFlagsCounts {
unsegmented: 0,
first: 8,
last: 8,
both: 2,
Expand All @@ -598,6 +643,7 @@ mod tests {
ordering_flags.insert(
Arc::clone(&rg_paired),
OrderingFlagsCounts {
unsegmented: 0,
first: 8,
last: 8,
both: 0,
Expand All @@ -607,6 +653,7 @@ mod tests {
ordering_flags.insert(
Arc::clone(&rg_single),
OrderingFlagsCounts {
unsegmented: 0,
first: 0,
last: 0,
both: 2,
Expand Down

0 comments on commit d5b5b5c

Please sign in to comment.