Skin cancer (MGI) · Cancer PRSweb

Scores Evaluation trait 172 Michigan Genomics Initiative

Skin cancer

Exact trait key: 172

GWAS round 2: Results shared 1st August 2018. Imputed genotypes from HRC plus UK10K & 1000 Genomes reference panels as released by UK Biobank in March 2018.

Binary MGI PRS-CS UKB GWAS (PHESANT) nominally significant

This PRS has a linked catalog record and a separate weight-file download.

Model details

PRS IDprsweb-phecode172-20001-1003-prs-cs-mgi-20200608

Legacy IDPRSWEB_PHECODE172_20001-1003_PRS-CS_MGI_20200608

Legacy modelPRS-CS

# SNPs1103220

Catalog record ID PGS000398

Discovery GWAS trait/source UKB GWAS (PHESANT)

Genome buildGRCh37/hg19

Tuning parameterauto

Reference Publication UKB_GWAS
PubMed 30104761

Last updated2020-06-08

Evaluation details

Evaluation cohortMichigan Genomics Initiative

Evaluation trait Cancer code, self-reported: skin cancer
172

Trait typeBinary

SexBoth

Cases3557

Controls14050

PRS performance

P value0.037

Pseudo-R²0.001

Brier score0.148

AAUC0.510

AAUC CI0.499-0.521

Percentile effect metricOR

Top percentile	OR	95% CI
1%	1.70	1.22-2.38
2%	1.44	1.12-1.84
5%	1.14	0.97-1.35
10%	1.10	0.97-1.24
25%	1.10	1.01-1.20

How to use this score page

Start with P value and the main performance metrics above to judge whether this model is useful in this cohort.
Use the top-percentile rows to see how strongly the score separates the upper PRS tail.
Use the linked PheWAS follow-up after the primary-trait evaluation, not as a substitute for it.

Flag guide: nominally significant passed the preserved release's nominal threshold; reversed effect means the observed direction ran opposite to the expected effect direction in the preserved evaluation.

Files and next steps

Start with the interactive PheWAS page to inspect notable associations for this score. Use the PRS PheWAS file for the full phenome scan, the exclusion PheWAS file for secondary associations after removing primary or related trait cases, and the weight link when you need the score construction itself. See the Method page for metric definitions and cohort context.