Remove founder & other small populations from default frequency sources

[julesjacobsen]

Exomiser 10 - 13.2.1 use the following default frequency sources:

frequencySources: [
  THOUSAND_GENOMES,
  TOPMED,
  UK10K,

  ESP_AFRICAN_AMERICAN, ESP_EUROPEAN_AMERICAN, ESP_ALL,

  EXAC_AFRICAN_INC_AFRICAN_AMERICAN, EXAC_AMERICAN,
  EXAC_SOUTH_ASIAN, EXAC_EAST_ASIAN,
  EXAC_FINNISH, EXAC_NON_FINNISH_EUROPEAN,
  EXAC_OTHER,

  GNOMAD_E_AFR,
  GNOMAD_E_AMR,
  #        GNOMAD_E_ASJ,
  GNOMAD_E_EAS,
  GNOMAD_E_FIN,
  GNOMAD_E_NFE,
  GNOMAD_E_OTH,
  GNOMAD_E_SAS,

  GNOMAD_G_AFR,
  GNOMAD_G_AMR,
  #        GNOMAD_G_ASJ,
  GNOMAD_G_EAS,
  GNOMAD_G_FIN,
  GNOMAD_G_NFE,
  GNOMAD_G_OTH,
  GNOMAD_G_SAS
]

The inclusion of the Finnish (FIN) and Other (OTH) populations should probably not be used as defaults as these are either founder populations (FIN) or too small a population (OTH) to give a reliable filtering frequency.

This is based on the ClinGen SVI BA1 recommendations and the gnomAD filtering allele frequency implementation.

From Updated recommendation for the benign stand-alone ACMG/AMP criterion

We have specified the data sets that we used for this analysis as the six defined subsets of the ExAC database (African, East Asian, European [non-Finnish], Latino, and South Asian)

and later

We have made clear that one should use this criterion to assign a variant as benign if its allele frequency is >0.05 in any one of the six specified data sets when there are at least 2,000 observed alleles at the site in question.

gnomAD filtering allele frequency

This calculation only includes non-bottlenecked populations. For gnomAD v2, popmax filtering allele frequency excludes Ashkenazi Jewish (asj), European Finnish (fin), and "Other" (oth) populations. For gnomAD v3, popmax filtering allele frequency excludes Amish (ami), Ashkenazi Jewish (asj), European Finnish (fin), Middle Eastern (mid), and "Other" (oth) populations.

[damiansm]

Good spot. Lets do the quick Pheval test and make it happen!

…

On Wed, Sep 6, 2023 at 5:17 PM Jules Jacobsen ***@***.***> wrote: Exomiser 10 - 13.2.1 use the following default frequency sources: frequencySources: [ THOUSAND_GENOMES, TOPMED, UK10K, ESP_AFRICAN_AMERICAN, ESP_EUROPEAN_AMERICAN, ESP_ALL, EXAC_AFRICAN_INC_AFRICAN_AMERICAN, EXAC_AMERICAN, EXAC_SOUTH_ASIAN, EXAC_EAST_ASIAN, EXAC_FINNISH, EXAC_NON_FINNISH_EUROPEAN, EXAC_OTHER, GNOMAD_E_AFR, GNOMAD_E_AMR, # GNOMAD_E_ASJ, GNOMAD_E_EAS, GNOMAD_E_FIN, GNOMAD_E_NFE, GNOMAD_E_OTH, GNOMAD_E_SAS, GNOMAD_G_AFR, GNOMAD_G_AMR, # GNOMAD_G_ASJ, GNOMAD_G_EAS, GNOMAD_G_FIN, GNOMAD_G_NFE, GNOMAD_G_OTH, GNOMAD_G_SAS] The inclusion of the Finnish (FIN) and Other (OTH) populations should probably not be used as defaults as these are either founder populations (FIN) or too small a population (OTH) to give a reliable filtering frequency. This is based on the ClinGen SVI BA1 recommendations and the gnomAD filtering allele frequency implementation. From Updated recommendation for the benign stand-alone ACMG/AMP criterion <https://pubmed.ncbi.nlm.nih.gov/30311383/> We have specified the data sets that we used for this analysis as the six defined subsets of the ExAC database (African, East Asian, European [non-Finnish], Latino, and South Asian) and later We have made clear that one should use this criterion to assign a variant as benign if its allele frequency is >0.05 in any one of the six specified data sets when there are at least 2,000 observed alleles at the site in question. gnomAD filtering allele frequency <https://gnomad.broadinstitute.org/help/faf> This calculation only includes non-bottlenecked populations. For gnomAD v2, popmax filtering allele frequency excludes Ashkenazi Jewish (asj), European Finnish (fin), and "Other" (oth) populations. For gnomAD v3, popmax filtering allele frequency excludes Amish (ami), Ashkenazi Jewish (asj), European Finnish (fin), Middle Eastern (mid), and "Other" (oth) populations. — Reply to this email directly, view it on GitHub <#513>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABHO4PAKN7F4VHGIVM5QLF3XZCOX3ANCNFSM6AAAAAA4NS2Q6U> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[julesjacobsen]

Results from GEL corpus of 742 samples with Exomiser 13.2.1 both with (default) and without populations (ASJ, FIN, OTH) filtering.

Result	Top	Top3	Top5	Top10	Found	MRR
Default	483	535	566	583	597	0.866
FIN,OTH	483	535	566	583	598	0.865

So this looks to be a benign change as we're gaining one extra case outside of the top 10.

Action

Remove ASJ, FIN, OTH populations from the default settings and examples.