A Breast Cancer Polygenic Risk Score Validation in 15,490 Brazilians using Exome Sequencing

A Breast Cancer Polygenic Risk Score Validation in 15,490 Brazilians using Exome Sequencing | medRxiv /* */ /* */ <!-- <!-- /*! * yepnope1.5.4 * (c) WTFPL, GPLv2 */ (function(a,b,c){function d(a){return"[object Function]"==o.call(a)}function e(a){return"string"==typeof a}function f(){}function g(a){return!a||"loaded"==a||"complete"==a||"uninitialized"==a}function h(){var a=p.shift();q=1,a?a.t?m(function(){("c"==a.t?B.injectCss:B.injectJs)(a.s,0,a.a,a.x,a.e,1)},0):(a(),h()):q=0}function i(a,c,d,e,f,i,j){function k(b){if(!o&&g(l.readyState)&&(u.r=o=1,!q&&h(),l.onload=l.onreadystatechange=null,b)){"img"!=a&&m(function(){t.removeChild(l)},50);for(var d in y[c])y[c].hasOwnProperty(d)&&y[c][d].onload()}}var j=j||B.errorTimeout,l=b.createElement(a),o=0,r=0,u={t:d,s:c,e:f,a:i,x:j};1===y[c]&&(r=1,y[c]=[]),"object"==a?l.data=c:(l.src=c,l.type=a),l.width=l.height="0",l.onerror=l.onload=l.onreadystatechange=function(){k.call(this,r)},p.splice(e,0,u),"img"!=a&&(r||2===y[c]?(t.insertBefore(l,s?null:n),m(k,j)):y[c].push(l))}function j(a,b,c,d,f){return q=0,b=b||"j",e(a)?i("c"==b?v:u,a,b,this.i++,c,d,f):(p.splice(this.i++,0,a),1==p.length&&h()),this}function k(){var a=B;return a.loader={load:j,i:0},a}var l=b.documentElement,m=a.setTimeout,n=b.getElementsByTagName("script")[0],o={}.toString,p=[],q=0,r="MozAppearance"in l.style,s=r&&!!b.createRange().compareNode,t=s?l:n.parentNode,l=a.opera&&"[object Opera]"==o.call(a.opera),l=!!b.attachEvent&&!l,u=r?"object":l?"script":"img",v=l?"script":u,w=Array.isArray||function(a){return"[object Array]"==o.call(a)},x=[],y={},z={timeout:function(a,b){return b.length&&(a.timeout=b[0]),a}},A,B;B=function(a){function b(a){var a=a.split("!"),b=x.length,c=a.pop(),d=a.length,c={url:c,origUrl:c,prefixes:a},e,f,g;for(f=0;f<d;f++)g=a[f].split("="),(e=z[g.shift()])&&(c=e(c,g));for(f=0;f<b;f++)c=x[f](c);return c}function g(a,e,f,g,h){var i=b(a),j=i.autoCallback;i.url.split(".").pop().split("?").shift(),i.bypass||(e&&(e=d(e)?e:e[a]||e[g]||e[a.split("/").pop().split("?")[0]]),i.instead?i.instead(a,e,f,g,h):(y[i.url]?i.noexec=!0:y[i.url]=1,f.load(i.url,i.forceCSS||!i.forceJS&&"css"==i.url.split(".").pop().split("?").shift()?"c":c,i.noexec,i.attrs,i.timeout),(d(e)||d(j))&&f.load(function(){k(),e&&e(i.origUrl,h,g),j&&j(i.origUrl,h,g),y[i.url]=2})))}function h(a,b){function c(a,c){if(a){if(e(a))c||(j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}),g(a,j,b,0,h);else if(Object(a)===a)for(n in m=function(){var b=0,c;for(c in a)a.hasOwnProperty(c)&&b++;return b}(),a)a.hasOwnProperty(n)&&(!c&&!--m&&(d(j)?j=function(){var a=[].slice.call(arguments);k.apply(this,a),l()}:j[n]=function(a){return function(){var b=[].slice.call(arguments);a&&a.apply(this,b),l()}}(k[n])),g(a[n],j,b,n,h))}else!c&&l()}var h=!!a.test,i=a.load||a.both,j=a.callback||f,k=j,l=a.complete||f,m,n;c(h?a.yep:a.nope,!!i),i&&c(i)}var i,j,l=this.yepnope.loader;if(e(a))g(a,0,l,0);else if(w(a))for(i=0;i (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0];var j=d.createElement(s);var dl=l!='dataLayer'?'&l='+l:'';j.src='//www.googletagmanager.com/gtm.js?id='+i+dl;j.type='text/javascript';j.async=true;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-P4HH5NV'); Skip to main content Home About Submit ALERTS / RSS Search for this keyword Advanced Search A Breast Cancer Polygenic Risk Score Validation in 15,490 Brazilians using Exome Sequencing View ORCID Profile Flávia Eichemberger Rius , View ORCID Profile Rodrigo Santa Cruz Guindalini , Danilo Viana , Júlia Salomão , Laila Gallo , Renata Freitas , Cláudia Bertolacini , Lucas Taniguti , Danilo Imparato , Flávia Antunes , Gabriel Sousa , View ORCID Profile Renan Achjian , Eric Fukuyama , Cleandra Gregório , Iuri Ventura , Juliana Gomes , Nathália Taniguti , View ORCID Profile Simone Maistro , José Eduardo Krieger , View ORCID Profile Yonglan Zheng , View ORCID Profile Dezheng Huo , View ORCID Profile Olufunmilayo I. Olopade , View ORCID Profile Maria Aparecida Koike , David Schlesinger doi: https://doi.org/10.1101/2024.04.21.24306089 Flávia Eichemberger Rius 1 Mendelics , São Paulo, Brazil 2 Comprehensive Center for Precision Oncology - C2PO, Centro de Investigação Translacional em Oncologia (CTO), Departamento de Radiologia e Oncologia, Instituto do Cancer do Estado de Sao Paulo (ICESP), Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo , Sao Paulo, Brazil PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Flávia Eichemberger Rius Rodrigo Santa Cruz Guindalini 3 Instituto D’Or de Pesquisa e Ensino (IDOR) , São Paulo, Brazil MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Rodrigo Santa Cruz Guindalini Danilo Viana 1 Mendelics , São Paulo, Brazil MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Júlia Salomão 1 Mendelics , São Paulo, Brazil MD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Laila Gallo 1 Mendelics , São Paulo, Brazil MD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Renata Freitas 1 Mendelics , São Paulo, Brazil PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Cláudia Bertolacini 1 Mendelics , São Paulo, Brazil PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Lucas Taniguti 1 Mendelics , São Paulo, Brazil MS Find this author on Google Scholar Find this author on PubMed Search for this author on this site Danilo Imparato 1 Mendelics , São Paulo, Brazil MS Find this author on Google Scholar Find this author on PubMed Search for this author on this site Flávia Antunes 1 Mendelics , São Paulo, Brazil Find this author on Google Scholar Find this author on PubMed Search for this author on this site Gabriel Sousa 1 Mendelics , São Paulo, Brazil Find this author on Google Scholar Find this author on PubMed Search for this author on this site Renan Achjian 1 Mendelics , São Paulo, Brazil Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Renan Achjian Eric Fukuyama 1 Mendelics , São Paulo, Brazil Find this author on Google Scholar Find this author on PubMed Search for this author on this site Cleandra Gregório 1 Mendelics , São Paulo, Brazil PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Iuri Ventura 1 Mendelics , São Paulo, Brazil PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Juliana Gomes 1 Mendelics , São Paulo, Brazil MS Find this author on Google Scholar Find this author on PubMed Search for this author on this site Nathália Taniguti 1 Mendelics , São Paulo, Brazil MS Find this author on Google Scholar Find this author on PubMed Search for this author on this site Simone Maistro 2 Comprehensive Center for Precision Oncology - C2PO, Centro de Investigação Translacional em Oncologia (CTO), Departamento de Radiologia e Oncologia, Instituto do Cancer do Estado de Sao Paulo (ICESP), Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo , Sao Paulo, Brazil PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Simone Maistro José Eduardo Krieger 4 Instituto do Coração, Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo - FMUSP , São Paulo, Brazil MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site Yonglan Zheng 5 Medicine and Human Genetics, Center for Clinical Cancer Genetics and Global Health, University of Chicago Medical Center , Chicago, USA PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Yonglan Zheng Dezheng Huo 6 Department of Public Health Sciences, University of Chicago , Chicago, USA MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Dezheng Huo Olufunmilayo I. Olopade 5 Medicine and Human Genetics, Center for Clinical Cancer Genetics and Global Health, University of Chicago Medical Center , Chicago, USA MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Olufunmilayo I. Olopade Maria Aparecida Koike 2 Comprehensive Center for Precision Oncology - C2PO, Centro de Investigação Translacional em Oncologia (CTO), Departamento de Radiologia e Oncologia, Instituto do Cancer do Estado de Sao Paulo (ICESP), Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo , Sao Paulo, Brazil MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site ORCID record for Maria Aparecida Koike David Schlesinger 1 Mendelics , São Paulo, Brazil MD, PhD Find this author on Google Scholar Find this author on PubMed Search for this author on this site For correspondence: david{at}mendelics.com.br Abstract Full Text Info/History Metrics Supplementary material Data/Code Preview PDF Abstract Background Brazil has a highly admixed population. Polygenic Risk Scores (PRS) have mostly been developed from European population studies, and their application to other populations is challenging. To assess the use of PRS for breast cancer (BC) risk in Brazil, we evaluated four PRSs in the Brazilian population. Methods We analyzed a Brazilian cohort composed of 6,206 women with a history of breast cancer and 8,878 unphenotyped adults as controls. Genomic variants were imputed from exomes and scores were calculated for all samples. Results After excluding individuals with known pathogenic or likely pathogenic variants in BRCA1 , BRCA2 , PALB2 , PTEN , or TP53 genes, and first-degree relatives of the probands, 5,598 cases and 8,767 controls remained. Four PRS models were compared, and PRS 3820 from Mavaddat et al. 2019 study showed the best performance, with an Odds Ratio (OR) of 1.43 per standard deviation increase (p-value: < 0.001) and an OR of 1.88 (p-value: < 0.001) for the top decile. PRS 3820 also performed well for different ancestry groups: East Asian majority (OR 1.59, p-value 0.004), Non-European majority (OR 1.45, p-value: <0.001), and European majority (OR: 1.43, p-value: <0.001). Conclusion Among different PRS, the PRS 3820 performed better in the highly admixed Brazilian population. This will allow a more precise BC risk assessment of mutation-negative women in Brazil. Introduction Breast cancer (BC) is a critical global health concern, representing the most common cancer diagnosed among women 1 . In Brazil, over 70,000 women are diagnosed with BC every year, accounting for 30% of all cancers in the female population 2 . Approximately 10% of all BC cases are attributable to germline pathogenic variants in susceptibility genes 3 . Rare variants in these genes account for approximately 25% of the genetic risk. The remaining genetic risk (∼75%) is derived from common, low penetrance variants that individually confer small risk, but which combined effect can be substantial 4 – 6 . Genome-wide association studies (GWASs) have been predominantly carried out in European populations 7 – 10 . Evaluation of PRS across different genetic and environmental backgrounds is essential to enable the implementation of genetic risk stratification strategies for individuals from non-European populations 11 . The Brazilian population exhibits a unique, highly admixed, genetic composition. It is mostly derived from a combination of Native Americans, Southern Europeans (Portuguese, Spanish, and Italian) that immigrated in the period 1500-1900, and Sub-Saharan Africans brought through extensive slave trading until the 1800s. More recently, from 1822 to the first half of the 1900s, other smaller waves of immigration also contributed to Brazil’s remarkable diversity, including Japanese, Lebanese, German, and Eastern Europeans 12 . Three in every four Brazilians have multiple genetic ancestries 13 , 14 . Given Brazil’s genetic diversity, any PRS developed in predominantly European populations requires validation before it can be used in clinical settings. Several laboratory methods are available for genotyping variants directly or indirectly (imputation), including microarrays, whole exome sequencing (WES), and whole genome sequencing (WGS). WES offers an affordable and scalable alternative to arrays and WGS, while allowing for simultaneous rare and common variant genotyping. In this study, we evaluate four BC PRSs 7 , 8 , 15 developed using WES in 15,490 Brazilians. Methods Study population A total of 15,490 individuals were selected for this study, including 6,362 women with breast cancer history, and 9,128 adult unphenotyped controls. Both clinical and genetic data were collected from a database of a College of American Pathology (CAP)-accredited laboratory (Mendelics, São Paulo, SP, Brazil). All BC and control subjects provided Informed Consent for use of retrospective anonymized data for research purposes. Samples were anonymized before analysis. Clinical records such as BC histological type and age of diagnosis were obtained from genetic test requisitions. The study was IRB-approved (CAAE: 70112423.3.0000.0068). Exome sequencing and imputation Exome sequencing data were generated from buccal swab or venous blood samples with standard protocol for Illumina Flex Exome Prep, using a custom probe set from Twist Biosciences. Sequencing was conducted in Illumina sequencers and the bioinformatics pipeline for data analysis followed Broad Institute’s GATK best practices ( https://gatk.broadinstitute.org/hc/en-us/sections/360007226651-Best-Practices-Workflo ws), with alignment to GRCh38. Imputation of exomes was based on a panel of 2,504 individuals of all ancestries from the 1000 Genomes Project (1KGP) 16 on GRCh38 (2017 release) ( https://www.internationalgenome.org/data-portal/data-collection/grch38 ). All regions captured from the exome sequencing comprehending at least 1x coverage, as well as off-target regions, were considered for the imputation, performed using Glimpse (v1.1.0) software 17 . Polygenic Risk Score Calculation Four BC PRSs with publicly available summary statistics, from three different studies, were evaluated in this work: Khera et al. 2018 7 , with 5,218 variants (PGS Catalog 18 , 19 ID: PGS000015); Mavaddat et al. 2019 8 PRSs (with 313 and 3,820 variants, respective PGS Catalog IDs: PGS000004 and PGS000007); and UK Biobank 15 (UKBB) PRS obtained from a variant thresholding (p-value < 10e-5) on summary statistics for phenotype code 20001_1002, with 7,538 variants. To overcome the constraints associated with variants derived from imputation, we filtered PRS variants based on their distance from exome bed kit and minor allele frequency (MAF). Variants with null betas (beta = 0) were removed from all PRSs. Additionally, the PRS from Mavaddat study, originally with 3,820 variants, had a pathogenic variant of moderate-penetrance in CHEK2 gene ( CHEK2 p.Ile157Thr - Clinvar: RCV000144596) which was removed to avoid conflation with monogenic risk. PRS calculation was performed using a software developed by Mendelics, evaluating the weighted sum of beta values, in which weights are based on the number of the individual’s alleles containing the variant of the PRS file. The sum is normalized by all beta positive and negative values so the final value can be between zero and one. Statistical analyses PRS values were standardized according to the control values prior to all statistical analyses. The effect size of PRS on breast cancer status was assessed using logistic regression, adjusting for z-scored PCs 1 to 10. AUC for the full dataset evaluation was obtained using the yardstick R package (yardstick.tidymodels.org/) roc_auc function, in the testing data split (25%). In order to find segmentation effect-sizes, individuals were classified into deciles or percentiles based on left-open and right-closed intervals. OR for deciles was calculated by first selecting only the decile analyzed and the median interval (40-60%) individuals as the control section, and binarizing it (0 = belongs to the control interval 40-60%, 1 = belongs to the decile analyzed, for example, 10%); and performing a logistic regression analysis on the binarized decile information with correction for PCs 1 to 10. A similar approach was conducted for calculating the OR on percentiles for comparison with Mavaddat’s 8 PRS validation. All comparisons with original studies were made with the values for the testing sets. For each ancestry proportion group, AUC was estimated using 10-fold cross-validation with the R package caret 20 . OR and CI for genes BRCA1 , BRCA2 , PALB2 , TP53 , ATM and CHEK2 , and for the variant R337H from TP53 were obtained using epitools R package 21 . All statistical tests performed were two-tailed. All analyses were performed in R version 4.4.2. Results Case-control sample selection and characteristics After removal of 211 subjects with a first-degree relationship, 73 with missing files necessary for imputation, and 122 with a low-quality imputation, a total of 15,084 subjects remained ( Supplementary Table 1 ). Individuals with P/LP variants in BC genes with OR > 5: BRCA1 , BRCA2 , TP53 , PALB2 , and PTEN were removed prior to PRS calculation (n = 629) ( Supplementary Methods ), resulting in a sample consisting of 5,598 women with a BC history, and 8,767 unphenotyped controls ( Table 1 ). View this table: View inline View popup Download powerpoint Table 1. Demographics of cases and controls in BC dataset used for PRS evaluation The ancestry composition of our admixed cohort, obtained using admixture 22 ( Supplementary Methods ), is shown in Figure 1 .The majority of individuals have EUR as their greatest ancestry proportion (median 84%, SD 18%). Significant fractions of AFR (median 6%, SD 12%) and AMR (median 8%, SD 7%) ancestries are observed, along with a variety of EUR proportions. A small proportion of EAS ancestry is also observed (median < 1%, SD 12%), primarily consisting of 214 individuals with over 70% of this ancestry. Download figure Open in new tab Figure 1. Ancestry composition of our Brazilian cohort. Estimated ancestries are shown as proportions per individual. Each thin bar represents one individual and their ancestry proportion. Europe (EUR) in purple, Africa (AFR) in blue, East Asia (EAS) in green and America (AMR) in yellow. Effect sizes of four different PRSs in the Brazilian population Four PRS files from three studies were selected for initial effect size investigation in our cohort: Broad 7 , 313 8 , 3820 8 and UKBB 15 ( Supplementary Table 2 ). All PRS files had their variants filtered to address only variants covered by the imputation of our exomes. PRSs were calculated for the exomes imputed into genomes (details described in the Methods ) and standardized to improve interpretability. Effects were corrected for the ten first PCs, and the results are reported in Supplementary Table 3 . Three of the four PRSs examined had statistically significant associations with breast cancer risk, with the ORs per SD ranging from 1.35 to 1.52 (PRS Broad : OR 1.52, 95% CI 1.46 - 1.59, AUC 0.614; PRS 3820 : OR 1.43, 95% CI 1.38 - 1.49, AUC 0.596; PRS 313 : OR 1.35, 95% CI 1.30 - 1.41, AUC 0.583). UKBB PRS was not significantly associated with breast cancer risk in our cohort (p value 0.40). Goodness of fit of the model was greater for PRS Broad (pseudo-R²: 0.062) and PRS 3820 (pseudo-R²: 0.054). Since PRS Broad , PRS 3820 and PRS 313 showed significant results per SD, they were used to split the data into deciles to evaluate BC risk conferred by PRS in each strata. These analyses were also corrected for the first ten PCs. In Figure 2 we can observe the staircase shape for all of the three PRSs. Especially the bottom and top deciles, the most critical when analyzing PRS data, show statistical significance (p < 0.001) and important effect sizes for all PRSs, ranging from 0.48 to 0.55 in the lowest decile, and 1.73 to 2.13 in the highest ( Supplementary Table 4 ). As the most notable result, PRS Broad shows over two-fold increased risk of breast cancer for women in the top decile (90-100%) compared to the middle decile (40-60%). Download figure Open in new tab Figure 2. Effect sizes by decile of PRS 313 , PRS 3820 and PRS Broad . Odds Ratios (OR) and Confidence Intervals (CI) are represented for PRS 313 (blue), PRS 3820 (orange) and PRS Broad (green). ORs for all PRS deciles were corrected for the first ten PCs. Deciles 5 and 6 were used as references to calculate ORs for the other deciles. Comparison of PRSs performances with the original studies The comparison of PRS metrics of this work with their original studies allows us to evaluate whether the metrics retain their accuracy and reliability in a genetically diverse and admixed population. Concerning OR per SD, all PRSs show a less pronounced value for our cohort compared to the original studies (OR PRS Broad : 1.52 vs. 1.56; PRS 3820 : 1.43 vs. 1.66, PRS 313 : 1.35 vs. 1.61). Similarly, the classification ability is less robust in our analysis (AUC Broad: 0.61 vs. 0.69; 3820: 0.60 vs. 0.64, 313: 0.58 vs. 0.63). This is an expected result given that our admixed population presents differences in allele frequencies and linkage disequilibrium, compared to the pure European populations used for both construction and validation of the PRSs. Considering OR per percentiles, we observed that the PRS 3820 shows a closer resemblance to Mavaddat’s results than PRS 313 ( Figure 3 ). This is primarily due to the top 1% OR, which shows a stronger effect for PRS 3820 than for PRS 313 (OR 2.93 vs. 1.98), aligning more closely with the original PRS 3820 result (OR 3.95) and deviating further from the original PRS 313 result (OR 4.04). Download figure Open in new tab Figure 3. Comparison of PRS 313 and PRS 3820 percentile effect sizes between original study and Brazilian cohort. (A) PRS 313 adjusted to this study (orange), with 283 variants, alongside the original from Mavaddat et al. study (green), with 313 variants. (B) PRS 3820 adapted in this study (orange), with 2,575 variants, alongside the original from Mavaddat et al. study (blue), with 3,820 variants. For PRS 3820 , the lower interval, comprehending the lowest 1% of PRS values, showed a smaller decrease in BC risk compared to the original study. This result is probably related to the small sample size of this section, with only 30 cases and 88 controls available to calculate OR. PRS evaluation by ancestry composition Since our sample contains a great majority of EUR ancestry proportion, we decided to evaluate the PRSs effect sizes in different ancestry compositions ( Figure 4A , Supplementary Methods ). All three bins had statistically significant (p < 0.001) ORs above 1.35 per SD for both PRS 313 and PRS 3820 , showing a positive association of the PRS value with increased BC risk for all ancestry groups ( Figure 4B ). PRS Broad shows significance for both European-related groups (p < 0.001) with the most prominent effect sizes among all PRSs (Non-European majority: OR 1.58, 95% CI 1.34 - 1.88; European majority: OR 1.52, 95% CI 1.47 - 1.58). Nevertheless, PRS Broad did not reach statistical significance for the East Asian majority group (p value = 0.08), indicating that this PRS may not be appropriate for use in the Brazilian admixed population. Download figure Open in new tab Figure 4. PRSs effect sizes by ancestry proportion. The cohort was split into three groups based on main ancestry: East Asian majority (>50% EAS), Non-European majority and European majority (A) Ancestry composition of each group, with colors representing continental ancestries for each subject: purple for EUR, blue for AFRm green for EAS and yellow for AMR. (B) Breast cancer ORs by PRS Broad , PRS 3820 and PRS 313 SD for the three ancestry groups. All p-values displayed were corrected for multiple-hypothesis testing using Bonferroni method. Correlation of PRS results derived from exomes and from genomes The subsequent analyses focused on PRS 3820 . A correlation of 0.74 (p value < 2.2e-16) was obtained between BC PRS 3820 values calculated from exomes followed by imputation and sequenced genomes, showing a concordance between both methods ( Supplementary Figure 1A ). Comparison of the proportion of individuals classified into deciles ( Supplementary Figure 1B ) shows that for the top and bottom deciles there is major concordance (57%) of the respective decile. Furthermore, the individuals classified in different deciles are mostly present in the surrounding deciles, indicating consistent results from both imputed and sequenced data. Breast cancer genes and PRS effect size comparison For the purpose of understanding how the PRS 3820 effect size compares to known high and moderate risk genes for BC, we have compared OR of the top PRS 3820 decile (PRS90) with all pathogenic variants located in TP53 , BRCA1 , BRCA2 , PALB2 , ATM and CHEK2 genes, plus the pathogenic variant R337H of TP53 gene ( Figure 5 ) in this cohort of individuals. Download figure Open in new tab Figure 5. Effect sizes of top decile of PRS and BC genes in BC risk. Effect sizes (OR and 95% CI) were obtained according to the presence of pathogenic variants in the genes TP53 , BRCA1 , BRCA2 , PALB2 , ATM and CHEK2 , or inclusion in the 90th to 100th percentiles of PRS 3820 (PRS90). As expected, TP53 , BRCA1 and BRCA2 present the most extreme BC risks ( TP53 OR: 14, 95% CI: 4.1-95; BRCA1 OR: 13.4, 95% CI: 9.2-20.3; BRCA2 OR: 8.8, 95% CI: 6.1-12.9, respectively). PRS90 risk (OR: 1.9, CI: 1.7-2.1) is slightly lower but comparable to the risk conferred by moderate risk BC genes ATM (OR: 2.5, CI: 1.7-3.9) and CHEK2 (OR: 2.1, CI: 1.5-3). Discussion In the present study we have validated two breast cancer PRSs developed from Europeans in the highly admixed Brazilian population. The PRSs adapted from Mavaddat et al. study with 283 (PRS 313 ) and 2,575 (PRS 3820 ) variants 8 showed statistically significant risk prediction values both per PRS SD and for the top decile compared with the middle deciles (p values < 0.001). PRS 3820 showed the best performance, with an OR of 1.43 per PRS SD (95% CI: 1.38-1.49) and 1.88 for the top decile (95% CI: 1.66 - 2.12). Furthermore, this PRS showed an OR per SD of 1.43 or above along different ancestry compositions (East Asian majority: OR 1.59, 95% CI 1.17-2.21, p value 0.004; Non-European majority: OR 1.45, 95% CI 1.24-1.71, p value < 0.001; and European majority: OR 1.43, 95% CI 1.38-1.48, p value < 0.001), highlighting its suitability for the diverse Brazilian population. The PRS with the best performance of this study is based on a previous study from Mavaddat et al. 2019, which developed and validated a PRS with 3,820 variants evaluating invasive BC risk. For all BC subtypes (ER+ and ER-) they found an OR of 1.71 per SD (CI: 1.64 1.79) in the validation set (n = 29,751; cases = 11,428), and OR 1.66 per SD (CI: 1.61 - 1.70) in the prospective set (n = 190,040; cases = 3,215). These values are even greater compared to the widely used 313 PRS (OR: 1.65 per SD; CI: 1.59 - 1.72 in validation set). However, they included a CHEK2 gene pathogenic variant in the PRS and worked with only invasive BC, which may have led to overestimating their OR values. A study from Liu and colleagues has evaluated another modification of the same PRS with 3,820 variants developed from Mavaddat et al. for African, Latin, and European populations 23 . According to the study, the effect size of this PRS in an European sample (n = 33,594) was 1.40 per SD, a result very similar to ours for a Brazilian sample (OR 1.43 per SD; n = 14,365). They deliberately have included women with in situ ductal BC as well as women with invasive BC, what they claim to be a reason for OR decline compared to the original study. Our study, however, does not distinguish BC types, therefore we hypothesize that both invasive and in situ BC are included, which may be a factor, together with genetic population structure, that decreased the OR compared to the original study. All of our PRS values were calculated according to a novel methodology: the imputation of exomes. This approach has demonstrated to be very successful for PRS calculation and assessment of BC risk in our study, and could be very interesting for laboratories that already perform exome sequencing as a cost-effective methodology to identify P/LP variants for BC. Multiple studies have compared low-pass genome sequencing with arrays for different applications, such as pharmacogenetics, GWAS and PRS calculation 24 , 25 , 26 . The study of Li et al. 25 reported improved accuracy for polygenic risk prediction of imputed low-pass genome compared to array imputation for both coronary artery disease and BC. Despite the slight difference we found between PRS values calculated from sequenced genomes and exomes with imputation (rho: 0.74), decile classification showed satisfactory concordance between both methods for the majority of results in the extreme deciles (1 and 10th), which are the most important to define decreased or increased risk. Unfortunately, it was not possible to assess the predictive power of PRS values calculated from genomes of BC patients due to unavailability of paired exome and genome data. Among familial BC cases, approximately 25% have a P/LP germline variant reported 27 . In the Brazilian population, a robust study with 1,663 breast cancer patients detected 20.1% of P/LP germline variants using multigene panel testing 4 , 5 . A 2017 study reported that 18% of the hereditary BC can be explained by a polygenic effect of variants discovered in a GWAS 28 . Therefore, employing this PRS in the clinical practice might bring an elucidation to BC Brazilian families without high or moderate-effect germline variants detected. Moreover, women without prior knowledge of their familial BC condition, or even those with a high PRS risk by chance, will have the possibility to be informed of their results and share them with their physicians to adopt preventive actions accordingly to their risk strata, such as intensifying surveillance adding breast magnetic resonance imaging to mammography screening 29 . In conclusion, we have validated both PRS 3820 and PRS 313 in the Brazilian population, demonstrating their potential utility for breast cancer polygenic risk assessment. Notably, PRS 3820 exhibited a greater effect size, with the top decile presenting a risk comparable to moderate-risk monogenic variants for BC. Future studies will be required to evaluate the combination of PRS with P/LP variants and clinical factors in order to deliver more informative results to patients, thus physicians can recommend prevention strategies based on their combined polygenic and monogenic BC risk. Ethics Statement This work was approved by the Ethics Committee of University of São Paulo - Faculdade de Medicina under the CAAE number 70112423.3.0000.0068. Funding Maria Aparecida Azevedo Koike Folgueira received research support from Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brazil (CNPq—308052/2022-6). Data Availability All variants and betas which compose the four evaluated PRSs are available as Supplementary Information. Individual cases and controls data are not publicly available due to the confidentiality consentment agreement signed by all included in the study. Competing Interests Flávia Eichemberger Rius, Danilo Viana, Júlia Salomão, Laila Gallo, Renata Freitas, Cláudia Bertolacini, Lucas Taniguti, Danilo Imparato, Flávia Antunes, Gabriel Sousa, Renan Achjian, Eric Fukuyama, Cleandra Gregório, Iuri Ventura, Juliana Gomes, Nathália Taniguti, and David Schlesinger are currently employed by Mendelics, or were employed at the time of the study. Rodrigo Guindalini acted as a consultant for AstraZeneca, Janssen Oncology, Roche/Genentech and Igenomix; received speaker honoraria from AstraZeneca, Bristol Myers Squibb, GlaxoSmithKline, Merck Sharpe & Dohme Brasil, Novartis, and Roche outside the submitted work; and has equity in Mendelics Análise Genômica. Olufunmilayo I. Olopade is co-founder at CancerIQ; serves as scientific advisor at Tempus; and has received research funding from Color Genomics and Roche/Genentech. José Eduardo Krieger, Yonglan Zheng, Dezheng Huo, Simone Maistro and Maria Aparecida Koike declare no competing interests. Author Contributions Generated Main Data Flávia Eichemberger Rius, Danilo Viana, Júlia Salomão, Laila Gallo, Renata Freitas, Cláudia Bertolacini, Lucas Taniguti, Danilo Imparato, Flávia Antunes, Gabriel Sousa, Renan Achjian, Eric Fukuyama, David Schlesinger. Analyzed Data Flávia Eichemberger Rius, Rodrigo Guindalini, Danilo Viana, Lucas Taniguti, Danilo Imparato, Flávia Antunes, Gabriel Sousa, Renan Achjian, Eric Fukuyama, Yonglan Zheng, Dezheng Huo, Olufunmilayo I. Olopade, Maria Aparecida Koike, David Schlesinger. Other Contributions Cleandra Gregório, Iuri Ventura, Juliana Gomes, Nathália Taniguti, Simone Maistro, José Eduardo Krieger. Acknowledgements We thank all individuals once sequenced in Mendelics laboratory who have consented to participate in this research. We also thank all UKBB participants for their contribution to the PRS hereby analyzed, and all authors from previous studies on BC PRSs in which we based our validation (Khera et al. 2018 and Mavaddat et al. 2019). Footnotes Results and figures updated after correction of PRS 313 variant set; results and figures updated after removal of 122 individuals with a low quality of imputation; all supplementary information related to the previous topics updated; formatting changed. References 1. ↵ Sung H , Ferlay J , Siegel RL , et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries . CA Cancer J Clin . 2021 ; 71 ( 3 ): 209 – 249 . doi: 10.3322/caac.21660 OpenUrl CrossRef PubMed 2. ↵ Instituto Nacional de Câncer. Estimativa 2023 : Incidência de Câncer No Brasil. (Ministério da Saúde, ed.). Ministério da Saúde; 2023 . 3. ↵ Nielsen FC , van Overeem Hansen T , Sørensen CS . Hereditary breast and ovarian cancer: new genes in confined pathways . Nat Rev Cancer . 2016 ; 16 ( 9 ): 599 – 612 . doi: 10.1038/nrc.2016.72 OpenUrl CrossRef PubMed 4. ↵ Melchor L , Benítez J . The complex genetic landscape of familial breast cancer . Hum Genet . 2013 ; 132 ( 8 ): 845 – 863 . doi: 10.1007/s00439-013-1299-y OpenUrl CrossRef PubMed 5. ↵ Guindalini RSC , Viana DV , Kitajima JPFW , et al. Detection of germline variants in Brazilian breast cancer patients using multigene panel testing . Sci Rep . 2022 ; 12 ( 1 ): 4190 . doi: 10.1038/s41598-022-07383-1 OpenUrl CrossRef PubMed 6. ↵ Shiovitz S , Korde LA . Genetics of breast cancer: a topic in evolution . Ann Oncol . 2015 ; 26 ( 7 ): 1291 – 1299 . doi: 10.1093/annonc/mdv022 OpenUrl CrossRef PubMed 7. ↵ Khera AV , Chaffin M , Aragam KG , et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations . Nat Genet . 2018 ; 50 ( 9 ): 1219 – 1224 . doi: 10.1038/s41588-018-0183-z OpenUrl CrossRef 8. ↵ Mavaddat N , Michailidou K , Dennis J , et al. Polygenic risk scores for prediction of breast cancer and breast cancer subtypes . Am J Hum Genet . 2019 ; 104 ( 1 ): 21 – 34 . doi: 10.1016/j.ajhg.2018.11.002 OpenUrl CrossRef PubMed 9. Zhang H , Ahearn TU , Lecarpentier J , et al. Genome-wide association study identifies 32 novel breast cancer susceptibility loci from overall and subtype-specific analyses . Nat Genet . 2020 ; 52 ( 6 ): 572 – 581 . doi: 10.1038/s41588-020-0609-2 OpenUrl CrossRef PubMed 10. ↵ Morra A , Escala-Garcia M , Beesley J , et al. Association of germline genetic variants with breast cancer-specific survival in patient subgroups defined by clinic-pathological variables related to tumor biology and type of systemic treatment . Breast Cancer Res . 2021 ; 23 ( 1 ): 86 . doi: 10.1186/s13058-021-01450-7 OpenUrl CrossRef PubMed 11. ↵ Mars N , Kerminen S , Feng Y-CA , et al. Genome-wide risk prediction of common diseases across ancestries in one million people . Cell Genomics . 2022 ; 2 ( 4 ):None. doi: 10.1016/j.xgen.2022.100118 OpenUrl CrossRef PubMed 12. ↵ Salzano , Freire-Maia FMN. As origens . In: Populações Brasileiras: Aspectos Demográficos, Genéticos e Antropológicos . ; 1967 . 13. ↵ Souza AM de , Resende SS , Sousa TN de , Brito CFA de . A systematic scoping review of the genetic ancestry of the Brazilian population . Genet Mol Biol . 2019 ; 42 ( 3 ): 495 – 508 . doi: 10.1590/1678-4685-GMB-2018-0076 OpenUrl CrossRef PubMed 14. ↵ Naslavsky MS , Scliar MO , Yamamoto GL , et al. Whole-genome sequencing of 1,171 elderly admixed individuals from São Paulo, Brazil . Nat Commun . 2022 ; 13 ( 1 ): 1004 . doi: 10.1038/s41467-022-28648-3 OpenUrl CrossRef PubMed 15. ↵ Sudlow C , Gallacher J , Allen N , et al. UK Biobank: an open access resource for identifying the causes of a wide range of complex diseases of middle and old age . PLoS Med . 2015 ; 12 ( 3 ): e1001779 . doi: 10.1371/journal.pmed.1001779 OpenUrl CrossRef PubMed 16. ↵ 1000 Genomes Project Consortium, Auton A , Brooks LD , et al. A global reference for human genetic variation . Nature . 2015 ; 526 ( 7571 ): 68 – 74 . doi: 10.1038/nature15393 OpenUrl CrossRef PubMed 17. ↵ Rubinacci S , Ribeiro DM , Hofmeister RJ , Delaneau O . Efficient phasing and imputation of low-coverage sequencing data using large reference panels . Nat Genet . 2021 ; 53 ( 1 ): 120 – 126 . doi: 10.1038/s41588-020-00756-0 OpenUrl CrossRef PubMed 18. ↵ Lambert SA , Gil L , Jupp S , et al. The Polygenic Score Catalog as an open database for reproducibility and systematic evaluation . Nat Genet . 2021 ; 53 ( 4 ): 420 – 425 . doi: 10.1038/s41588-021-00783-5 OpenUrl CrossRef PubMed 19. ↵ Lambert SA , Wingfield B , Gibson JT , et al. Enhancing the Polygenic Score Catalog with tools for score calculation and ancestry normalization . Nat Genet . 2024 ; 56 ( 10 ): 1989 – 1994 . doi: 10.1038/s41588-024-01937-x OpenUrl CrossRef 20. ↵ Kuhn M . Building Predictive Models in R Using the caret Package . J Stat Softw . 2008 ; 28 ( 5 ). doi: 10.18637/jss.v028.i05 OpenUrl CrossRef PubMed 21. ↵ Aragon TJ , Fay MP , Wollschlaeger D , Omidpanah A. Epitools: Epidemiology Tools. Tools for Training and Practicing Epidemiologists Including Methods for Two-Way and Multi-Way Contingency Tables . CRAN ; 2020 . 22. ↵ Alexander DH , Novembre J , Lange K . Fast model-based estimation of ancestry in unrelated individuals . Genome Res . 2009 ; 19 ( 9 ): 1655 – 1664 . doi: 10.1101/gr.094052.109 OpenUrl Abstract / FREE Full Text 23. ↵ Liu C , Zeinomar N , Chung WK , et al. Generalizability of polygenic risk scores for breast cancer among women with european, african, and latinx ancestry . JAMA Netw Open . 2021 ; 4 ( 8 ): e2119084 . doi: 10.1001/jamanetworkopen.2021.19084 OpenUrl CrossRef 24. ↵ Wasik K , Berisa T , Pickrell JK , et al. Comparing low-pass sequencing and genotyping for trait mapping in pharmacogenetics . BMC Genomics . 2021 ; 22 ( 1 ): 197 . doi: 10.1186/s12864-021-07508-2 OpenUrl CrossRef PubMed 25. ↵ Li JH , Mazur CA , Berisa T , Pickrell JK . Low-pass sequencing increases the power of GWAS and decreases measurement error of polygenic risk scores compared to genotyping arrays . Genome Res . 2021 ; 31 ( 4 ): 529 – 537 . doi: 10.1101/gr.266486.120 OpenUrl Abstract / FREE Full Text 26. ↵ Chaubey A , Shenoy S , Mathur A , et al. Low-Pass Genome Sequencing: Validation and Diagnostic Utility from 409 Clinical Cases of Low-Pass Genome Sequencing for the Detection of Copy Number Variants to Replace Constitutional Microarray . J Mol Diagn . 2020 ; 22 ( 6 ): 823 – 840 . doi: 10.1016/j.jmoldx.2020.03.008 OpenUrl CrossRef PubMed 27. ↵ Bahcall O . Common variation and heritability estimates for breast, ovarian and prostate cancers . Nat Genet. January 1 , 2019 . doi: 10.1038/ngicogs.1 OpenUrl CrossRef 28. ↵ Michailidou K , Lindström S , Dennis J , et al. Association analysis identifies 65 new breast cancer risk loci . Nature . 2017 ; 551 ( 7678 ): 92 – 94 . doi: 10.1038/nature24284 OpenUrl CrossRef PubMed 29. ↵ Monticciolo DL , Newell MS , Moy L , Lee CS , Destounis SV . Breast Cancer Screening for Women at Higher-Than-Average Risk: Updated Recommendations From the ACR . J Am Coll Radiol . 2023 ; 20 ( 9 ): 902 – 914 . doi: 10.1016/j.jacr.2023.04.002 OpenUrl CrossRef View the discussion thread. Back to top Previous Next Posted January 27, 2025. Download PDF Supplementary Material Data/Code Email Thank you for your interest in spreading the word about medRxiv. NOTE: Your email address is requested solely to identify you as the sender of this article. Your Email * Your Name * Send To * Enter multiple addresses on separate lines or separate them with commas. You are going to email the following A Breast Cancer Polygenic Risk Score Validation in 15,490 Brazilians using Exome Sequencing Message Subject (Your Name) has forwarded a page to you from medRxiv Message Body (Your Name) thought you would like to see this page from the medRxiv website. Your Personal Message CAPTCHA This question is for testing whether or not you are a human visitor and to prevent automated spam submissions. Share A Breast Cancer Polygenic Risk Score Validation in 15,490 Brazilians using Exome Sequencing Flávia Eichemberger Rius , Rodrigo Santa Cruz Guindalini , Danilo Viana , Júlia Salomão , Laila Gallo , Renata Freitas , Cláudia Bertolacini , Lucas Taniguti , Danilo Imparato , Flávia Antunes , Gabriel Sousa , Renan Achjian , Eric Fukuyama , Cleandra Gregório , Iuri Ventura , Juliana Gomes , Nathália Taniguti , Simone Maistro , José Eduardo Krieger , Yonglan Zheng , Dezheng Huo , Olufunmilayo I. Olopade , Maria Aparecida Koike , David Schlesinger medRxiv 2024.04.21.24306089; doi: https://doi.org/10.1101/2024.04.21.24306089 Share This Article: Copy Citation Tools A Breast Cancer Polygenic Risk Score Validation in 15,490 Brazilians using Exome Sequencing Flávia Eichemberger Rius , Rodrigo Santa Cruz Guindalini , Danilo Viana , Júlia Salomão , Laila Gallo , Renata Freitas , Cláudia Bertolacini , Lucas Taniguti , Danilo Imparato , Flávia Antunes , Gabriel Sousa , Renan Achjian , Eric Fukuyama , Cleandra Gregório , Iuri Ventura , Juliana Gomes , Nathália Taniguti , Simone Maistro , José Eduardo Krieger , Yonglan Zheng , Dezheng Huo , Olufunmilayo I. Olopade , Maria Aparecida Koike , David Schlesinger medRxiv 2024.04.21.24306089; doi: https://doi.org/10.1101/2024.04.21.24306089 Citation Manager Formats BibTeX Bookends EasyBib EndNote (tagged) EndNote 8 (xml) Medlars Mendeley Papers RefWorks Tagged Ref Manager RIS Zotero Tweet Widget Facebook Like Google Plus One Subject Area Genetic and Genomic Medicine Subject Areas All Articles Addiction Medicine (574) Allergy and Immunology (865) Anesthesia (304) Cardiovascular Medicine (4460) Dentistry and Oral Medicine (445) Dermatology (383) Emergency Medicine (611) Endocrinology (including Diabetes Mellitus and Metabolic Disease) (1517) Epidemiology (15251) Forensic Medicine (31) Gastroenterology (1132) Genetic and Genomic Medicine (6621) Geriatric Medicine (669) Health Economics (1002) Health Informatics (4564) Health Policy (1372) Health Systems and Quality Improvement (1617) Hematology (544) HIV/AIDS (1272) Infectious Diseases (except HIV/AIDS) (15938) Intensive Care and Critical Care Medicine (1107) Medical Education (624) Medical Ethics (147) Nephrology (670) Neurology (6642) Nursing (346) Nutrition (1001) Obstetrics and Gynecology (1148) Occupational and Environmental Health (957) Oncology (3350) Ophthalmology (981) Orthopedics (369) Otolaryngology (421) Pain Medicine (436) Palliative Medicine (130) Pathology (665) Pediatrics (1698) Pharmacology and Therapeutics (694) Primary Care Research (714) Psychiatry and Clinical Psychology (5464) Public and Global Health (9259) Radiology and Imaging (2212) Rehabilitation Medicine and Physical Therapy (1372) Respiratory Medicine (1198) Rheumatology (598) Sexual and Reproductive Health (716) Sports Medicine (533) Surgery (715) Toxicology (100) Transplantation (289) Urology (265) (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a038f1005e1458f4',t:'MTc4MDA5MzgwMw=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();