Complex SV intervals support #8521

mwalker174 · 2023-09-15T17:50:37Z

Complex (CPX) SV records no longer use END2/CHR2 INFO fields and instead are defined by CPX_INTERVALS in addition to normal coordinates.
Adds new class SVCallRecord.ComplexEventInterval for storing complex event intervals, which are associated with a particular SV type.
SV clustering now takes into account CPX subtype and intervals, requiring intervals to match on type and interval similarity as defined by the clustering parameters (e.g. minimum reciprocal overlap).
Improved reciprocal translocation (CTX) record handling. In particular, these records are now treated similarly to BNDs for clustering purposes.
Added unit tests for new CPX cases.

epiercehoffman

Thanks for working on this - complex SVs make for complex code. My primary hesitation is related to the way length is calculated for CPX events - let me know what you think. There are some other smaller comments/questions as well, and you'll need to rebase as I anticipate conflicts with my recent CTX PR.

src/main/java/org/broadinstitute/hellbender/tools/sv/SVCallRecord.java

src/main/java/org/broadinstitute/hellbender/tools/sv/SVCallRecordUtils.java

epiercehoffman · 2024-06-27T20:08:37Z

src/main/java/org/broadinstitute/hellbender/tools/sv/SVCallRecordUtils.java

        final Integer end2;
        final String chr2;
        if (type == GATKSVVCFConstants.StructuralVariantAnnotationType.BND
-            || type == GATKSVVCFConstants.StructuralVariantAnnotationType.CTX) {
+                || type == GATKSVVCFConstants.StructuralVariantAnnotationType.CTX) {
+            end = record.getPositionA();


Note for future: To handle complex translocations, there need to be breakpoints at CHROM:POS, CHROM:END, CHR2:POS2 (which we currently don't have), and CHR2:END2.

Made a TODO note. Would it also be possible to use the CPX intervals?

Just looked back at my conversation with Xuefang about this - for annotation we discussed that we should check CPX_INTERVALS for complex CTX, but in her CTX_INV example only the INV was in CPX_INTERVALS. For the CTX we used the other positions. If the INV is on the first chromosome, then the CTX breakpoints on that chromosome are defined by CHROM:POS and CHROM:END. But if the INV is on the second chromosome, then the CTX breakpoints are not fully defined by CHR2:END2 - there would need to be a CHR2:POS2.

epiercehoffman · 2024-07-01T19:39:06Z

src/main/java/org/broadinstitute/hellbender/tools/sv/SVCallRecord.java

+        private final GATKSVVCFConstants.StructuralVariantAnnotationType intervalType;
+        private final SimpleInterval interval;
+
+        public ComplexEventInterval(final String str) {


I was thinking about whether it would make sense to use this in SVAnnotate to harmonize across the different SV tools more. But in SVAnnotate I treat complex SV intervals the same way as any SV interval so that the annotation methods can be interchangeable. Maybe it could be done with some class hierarchy though? Or you could use SVSegment? The main difference is the class method to parse the interval string. I didn't come to a conclusion but wanted to put the thought out there of unifying the SV methods more if it's possible

This is a great point. This class is completely redundant with SVSegment, so I've moved that to its own class. However, I've made ComplexEventInterval a subclass with encode()/decode() methods. We could delete the new class altogether, but I think we should have this separate class since each is used in different tools. This will make it easier to make independent changes later if we need to.

src/main/java/org/broadinstitute/hellbender/tools/sv/SVCallRecordUtils.java

src/main/java/org/broadinstitute/hellbender/tools/sv/cluster/CanonicalSVLinkage.java

src/test/java/org/broadinstitute/hellbender/tools/sv/cluster/SVClusterEngineTest.java

src/main/java/org/broadinstitute/hellbender/tools/sv/cluster/CanonicalSVLinkage.java

src/test/java/org/broadinstitute/hellbender/tools/sv/cluster/SVClusterEngineTest.java

Cpx clustering support with testing Fix complex subtype check Complex event intervals Fix typos Another typo Address reviewer comments

gatk-bot · 2024-07-09T16:03:39Z

Github actions tests reported job failures from actions build 9859920680
Failures in the following jobs:

Test Type	JDK	Job ID	Logs
integration	17.0.6+10	9859920680.11	logs
integration	17.0.6+10	9859920680.0	logs

epiercehoffman · 2024-07-10T21:03:26Z

src/main/java/org/broadinstitute/hellbender/tools/spark/sv/utils/GATKSVVCFConstants.java

+    public static final Set<ComplexVariantSubtype> COMPLEX_POINT_SUBTYPES =
+            Collections.set(
+                    GATKSVVCFConstants.ComplexVariantSubtype.dDUP,
+                    GATKSVVCFConstants.ComplexVariantSubtype.dDUP_iDEL


dDUP_iDEL events don't have POS = END because there is a deletion at the sink

I couldn't actually find where this constant gets used though?

Whoops. I was considering this but didn't end up using it

src/test/java/org/broadinstitute/hellbender/tools/sv/cluster/SVClusterEngineTest.java

epiercehoffman

Thanks for making those changes!

epiercehoffman requested changes Jul 1, 2024

View reviewed changes

Improve ctx handling; remove end2/chr2 from cpx

e906b51

Cpx clustering support with testing Fix complex subtype check Complex event intervals Fix typos Another typo Address reviewer comments

mwalker174 force-pushed the mw_cpx_end branch from a1d2998 to e906b51 Compare July 9, 2024 15:29

Do not assume sorted complex intervals

5b60b2e

epiercehoffman reviewed Jul 10, 2024

View reviewed changes

epiercehoffman approved these changes Jul 10, 2024

View reviewed changes

Undo COMPLEX_POINT_SUBTYPES

8c04685

mwalker174 merged commit 23a38ce into master Jul 11, 2024
20 checks passed

mwalker174 deleted the mw_cpx_end branch July 11, 2024 19:57

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Complex SV intervals support #8521

Complex SV intervals support #8521

mwalker174 commented Sep 15, 2023

epiercehoffman left a comment

epiercehoffman Jun 27, 2024

mwalker174 Jul 8, 2024

epiercehoffman Jul 8, 2024

epiercehoffman Jul 1, 2024

mwalker174 Jul 8, 2024

gatk-bot commented Jul 9, 2024 •

edited

Loading

epiercehoffman Jul 10, 2024

epiercehoffman Jul 10, 2024

mwalker174 Jul 11, 2024

epiercehoffman left a comment

Complex SV intervals support #8521

Complex SV intervals support #8521

Conversation

mwalker174 commented Sep 15, 2023

epiercehoffman left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gatk-bot commented Jul 9, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

epiercehoffman left a comment

Choose a reason for hiding this comment

gatk-bot commented Jul 9, 2024 •

edited

Loading