We should add a list that points to references that supports the repeat and/or the levels of pathogenicity. References could be urls to for example OMIM or pubmed. What do you think of that @dnil ? This would make the whole thing more transparent and trackable. I'll be happy to help out with this of course.

hi,

an issue I just realized we've had for a while. When running stranger on expansionhunter (3.0 and 4.0 makes no difference) we get the below posted errors. We use stranger 0.8 and the latest hg38 catalogue from the repo. What I don't really understand, for instance, expansionhunter does two separate call for huntingtons LocusID HTT, with two different repeat units. But stranger only annotates one of these, the other one throws the error: WARNING No info for repeat id HTT_CCG. Is there something I am missing?

2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id DAB1_ATTTT1
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id DAB1_ATTTC
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id DAB1_ATTTT2
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id ATXN7_GCC
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id CNBP_CAGA
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id CNBP_CA
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id HTT_CCG
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id RAPGEF2_TTTTA_5P
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id RAPGEF2_TTTTA_3P
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id SAMD12_TTTTA_3P
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id SAMD12_TTTTA_5P
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id FXN_A
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id ATXN8OS_CTA
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id POLG_I
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id POLG_II
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id TNRC6A_TTTTA_5P
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id TNRC6A_TTTTA_3P
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id EIF4A3_CACA20
2021-06-09 10:30:46 rs-n1 stranger.utils[65859] WARNING No info for repeat id NOP56_CGCCTG

STRUC definitions are pretty crucial for some loci, and we do have support to flag some of the component () as pathogenic.

It would also be useful with a conversion so we can make the target beds from json and not have to keep them synced manually.

Somehow mistakes have been made for the associations between gene names (LocusId) and HGNCid. We discovered this in Lund when a gene panel contained only the gene CBL but the STR module displayed results for both CBL and BEAN1. At a quick review I notice that BEAN1 and CBL both have assigned HGNCId 1541 (correct for CBL but not for BEAN1). Further down in the stranger list CSTB and DAB1 both have 2482 (correct for CSTB but not for DAB1). It's possible that there are more discrepancies, I haven't done a systematic check, so please review the entire catalog.

As a downstream user, loading samples into Scout, I would like to see Stranger produce a simple ranking, where pathogenic expansions in known disease loci appear on top, pre-mutations next perhaps with a slight bonus to variants near the pathogenic limit, followed by normal variants and last additional loci of potential interest.

Hi !

Could you, please, add recently described repeat expansion disorders:

1. SCA4 https://pubmed.ncbi.nlm.nih.gov/38035881/
2. SCA27B https://pubmed.ncbi.nlm.nih.gov/37165652/
   Both of them appear to be quite common.

Would be nifty to tag inheritance models followed, so that they could be matched to disorder. Compare genmod models, but aware of expansion status (full_mutation / intermediate) rather than just existence of two variants.

New disease loci, new definitions needed. See Clinical-Genomics/reference-files#6. And we should merge the definition json and stranger resource json to make keeping them up to date as easy as can be.

Also remove some debug logging probably.

I think the POLG hg38 coordinates might be wrong here:

https://github.com/moonso/stranger/blob/master/stranger/resources/variant_catalog_grch38.json#L1220-L1221

In my own records the CTG repeat is @ chr15:89333590-89333628

Is your feature request related to a problem? Please describe.
I'm always frustrated when woke test fails on the branch name.

Describe the solution you'd like
Rename master branch to main.

Accommodate PacBio TRGT caller VCFs and annotate.

Disease mode of inheritance and Clinical repeat unit

TRGT VCFs have biallelic sites, and important per allele info in the FORMAT tags, e.g. AL and MC. We need to decompose

Tested bcftools norm and vt decompose -s but they can't be bothered with non-standard FORMAT fields, only decomposing FORMAT.DP and the like.

Hi!

I've noticed that after updating the demo variant catalog for hg38, many reference region coordinates and/or locus structures have changed. Interestingly, some of these coordinates also differ from those in the newest version of the original ExpansionHunter variant catalog. Some of the affected loci IDs include ATN1, ATXN1, ATXN3, PHOX2B, TBP, FOXL2, HOXD13, RUNX2, ZIC2, ZIC3. Is this intentional?

There appear to be conditions under which ExHu v3 reverts to the RC-style REPID naming, although we haven't quite figured why yet. Enable both and be done with it.

I thought I had solved this in #59, but I think when I changed line 260 from continue to repeat_res.extend([0]) the error consisted...

Could you have a look? HG002_Revio.vcf.gz

2024-05-31 14:40:51 stuart stranger.cli[41] INFO Running stranger version 0.9.0
  2024-05-31 14:40:51 stuart stranger.cli[41] INFO Parsing repeats file variant_catalog_grch38.json
  2024-05-31 14:40:51 stuart stranger.cli[41] INFO Vcf is zipped
  Traceback (most recent call last):
    File "/usr/local/bin/stranger", line 10, in <module>
      sys.exit(base_command())
               ^^^^^^^^^^^^^^
    File "/usr/local/lib/python3.12/site-packages/click/core.py", line 1157, in __call__
      return self.main(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/usr/local/lib/python3.12/site-packages/click/core.py", line 1078, in main
      rv = self.invoke(ctx)
           ^^^^^^^^^^^^^^^^
    File "/usr/local/lib/python3.12/site-packages/click/core.py", line 1434, in invoke
      return ctx.invoke(self.callback, **ctx.params)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/usr/local/lib/python3.12/site-packages/click/core.py", line 783, in invoke
      return __callback(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/usr/local/lib/python3.12/site-packages/click/decorators.py", line 33, in new_func
      return f(get_current_context(), *args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/usr/local/lib/python3.12/site-packages/stranger/cli.py", line 155, in cli
      repeat_data = get_repeat_info(variant_info, repeat_information)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/usr/local/lib/python3.12/site-packages/stranger/utils.py", line 209, in get_repeat_info
      repeat_res = get_trgt_repeat_res(variant_info, repeat_info)
                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/usr/local/lib/python3.12/site-packages/stranger/utils.py", line 269, in get_trgt_repeat_res
      pathologic_counts = int(allele)
                          ^^^^^^^^^^^
  ValueError: invalid literal for int() with base 10: '.'

Not entirely sure we got that right with the normal allele if there are more than one alt to play with..

• Incorporate new REPIDs (include coordinates)
• Change tsv file to json

Compare https://github.com/Illumina/ExpansionHunter/pull/78/files!
Remember to sync with https://github.com/Clinical-Genomics/reference-files.

Eg the readme is lagging a release or three behind..

A few more details that might need some clarification:

These 4 loci have different NormalMax values in 37 vs. 38 (and ATN1 also has different PathologicMin):

ATN1 35 34
DMPK 34 37
FMR1 55 65
TBP 40 44

These loci have differing LocusStructure (even though DisplayRU is the same):

TBP (CAN)* (GCA)*
NIPA1 (NGC)* (GCG)*

TBP locus GRCh38 coordinates match the EHv4 catalog coordinates (https://github.com/Illumina/ExpansionHunter/blob/master/variant_catalog/hg38/variant_catalog.json#L499)
but the GRCh37 coordinates don't match the EHv4 catalog (https://github.com/Illumina/ExpansionHunter/blob/master/variant_catalog/hg19/variant_catalog.json#L260)

The situation is the same for NIPA1.

Hello.
I see some commits mentioning version 0.8.1 are pushed 11 days ago, but no new release was published here: https://github.com/Clinical-Genomics/stranger/releases

Is release missing or still not completed yet?

Thank you for making these variant catalogs publicly available!
I think the GLS GRCh37 coordinates should be 2:191745598-191745646 based on liftover from GRCh38

Add HOXD13, HOXA13, FOXL2, LRP12, MUC1, EIF4A3, ...
See https://www.ncbi.nlm.nih.gov/books/NBK535148/

Seems the v3rc naming style changed again. Check and adapt!
Now IDs like NOP56_CGCCTG instead of chr coords.

E.g. INFO.REF, describing the count/MC expected for the reference locus.

GeneReviews https://www.ncbi.nlm.nih.gov/books/NBK535148/ and others.
Two tags: pubmed id, gene review id.

Similarly, the start coordinate for RFC1 appears to be 1 off on GRCh37.
The 0-based coordinates should be chr4:39,350,044-39,350,099.

Make sure Stranger and reference files use HGNC-ids that accurately reflect or at least can be tied to corresponding diseases (especially for filtering in Scout).

It is straighforward to produce normal population calls for each locus. These should be employed to show the rarity of the present expansions size, e g as a quantile value for the less selected population.

Would it be possible to annotate a multi-sample VCF?

Or if not, give an error if a multi-sample VCF is supplied 🙂

Hi,
in the latest version of the variant catalog there are three different variantIds for DAB1; DAB1_ATTTT1, DAB1_ATTTC and DAB1_ATTTT2. DAB1_ATTTC is the clinically significant one but this is not recognised by scout as the status field is left blank. For other variants with more than one variantId there seems to be the rule to give the clinically significant variant the same name as given for the LocusId. For instance, ATXN7 has two variantId; ATXN7 and ATXN7_GCC. When uploaded to scout they all seem to get a status annotation for the most significant variant in multiple variantId loci.

It might be a reason for the naming of DAB1 variantIds in the catalog but if not, it's probably an honest mistake?

Kind Regards,
Paul