Germline mutational landscape in a large cohort of Indian breast cancer

Abstract

Breast cancer has emerged as the leading malignancy among Indian women, with a notable subset of cases linked to inherited genetic predisposition. Germline testing is a vital tool for identifying individuals at elevated risk, guiding targeted therapies such as PARP inhibitors, and informing preventive strategies through family-based screening. However, population-specific data from India remain limited, and most studies have focused narrowly on BRCA1/2. In this study, we employed high-depth germline whole exome sequencing to comprehensively evaluate cancer susceptibility in 479 unselected Indian breast cancer patients sourced from a National Cancer Tissue Biobank. Pathogenic or likely pathogenic variants were identified in 24.6% of the cohort, including 8.35% in BRCA1/2. Notably, 67% of these findings were in non-BRCA genes, including HRR and tumor suppressor genes, which would be missed by BRCA-only testing. The unbiased, broad-panel approach enabled evaluation of emerging susceptibility genes such as RECQL. Additionally, we identified two clinically significant DPYD variants (c.1679T > G and c.1905 + 1G > A), associated with fluoropyrimidine toxicity risk. Beyond cancer risk, medically actionable secondary findings were detected in 21.7% of individuals, with variants in non-oncology genes listed in ACMG SF v3.2. Together, these findings highlight the genetic heterogeneity of hereditary breast cancer in India and demonstrate the clinical utility of comprehensive germline testing. While targeted multi-gene panels may suffice in clinical settings, whole-exome sequencing provides additional benefits by uncovering ancestry-specific susceptibility genes and actionable variants relevant for therapy and prevention in genetically diverse populations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-025-15360-w.

Introduction

Breast cancer has become the most frequently diagnosed malignancy among Indian women, surpassing cervical cancer and rising from the fourth most common female cancer in the 1990 s to the leading position today [1]. As per GLOBOCAN 2022, India reports ~ 192,020 new breast cancer and 47,333 ovarian cancer cases annually [2]. While this trend reflects changes in lifestyle, reproductive behavior, and awareness, a notable subset, particularly with early onset or family history, is driven by hereditary alterations. Over 10% of breast and ovarian cancer patients carry pathogenic germline mutations in high-penetrance genes such as BRCA1 and BRCA2³, which confer lifetime cancer risks of up to 69–72% for breast and 17–44% for ovarian cancer [3].

Despite advances in detection and treatment, germline pathogenic variants in high-penetrance genes such as BRCA1, BRCA2, and those in DNA damage repair pathways remain major contributors to breast cancer risk. Beyond risk assessment, BRCA1/2 and other HRR gene mutations also serve as prognostic and predictive biomarkers, conferring increased sensitivity to platinum-based chemotherapy, particularly in triple-negative breast cancer (TNBC) and high-grade serous ovarian cancers, and predicting strong responses to PARP inhibitors [4, 5].

A population-scale analysis of IndiGen whole-genome data from 1,029 healthy Indians estimated pathogenic germline variant prevalence at 1 in 342 for BRCA1 and 1 in 256 for BRCA2, underscoring a significant carrier burden and the need for ancestry-specific interpretation frameworks [6]. Patient-based studies report widely varying BRCA mutation prevalence (2.9–28%) depending on cohort and methodology [7, 8]. A recent multicentric study for, instance, found 25.2% prevalence among Indian ovarian cancer patients, irrespective of family history [9]. The contribution of non-BRCA genes to HBOC remains underreported, partly due to reliance on PCR assays that miss structural, rare, and novel variants.

NGS has transformed hereditary cancer testing with greater depth and coverage than PCR yet its uptake in India is limited and large-scale germline data remain scarce [10]. Whole-exome sequencing (WES) offers broad detection of known and novel susceptibility variants across pathways beyond BRCA1/2 [11], while also identifying pharmacogenomic variants and secondary findings relevant to treatment and prevention. In India’s genetically diverse and understudied population, such comprehensive approaches are essential for refining risk assessment and enabling personalized care.

In this study, we analyzed germline WES data from 479 unselected breast cancer patients, with biospecimens and clinical data sourced from the National Cancer Tissue Biobank (NCTB) at the Indian Institute of Technology Madras (IITM). This represents one of the largest germline sequencing efforts in Indian breast cancer patients to date. Using ACMG/AMP-guided variant classification, we aimed to (i) delineate pathogenic and likely pathogenic variants in BRCA1/2 and other cancer susceptibility genes, (ii) assess population specificity and recurrence, including in understudied genes such as RECQL, and (iii) identify the prevalence of clinically actionable pharmacogenomic markers and secondary findings.

Materials and methods

Patient samples

In this study, we analyzed a comprehensive collection of 479 frozen tissue samples along with corresponding sample data, sourced from the NCTB at IITM. The collection process was rigorously conducted after the Indian Institute of Technology Madras, Chennai Institutional Ethical Committee approvals (IITM/IEC/2014043; IEC/2019-03/SM/01/01). All patients gave written informed consent regarding the collection and storage of cancer tissue samples and the use of these samples for research purposes and publications. For each case, paired tumor and adjacent normal tissue samples were collected. The normal counterparts were independently reviewed and confirmed by an experienced pathologist to ensure the absence of tumor infiltration and to validate them as true normal controls. The study was conducted in accordance with the 1964 Helsinki Declaration and later amendments. The median age of the present cohort is ~ 57 years, with a lower age of 28 years and an upper 85 years.

DNA isolation and library preparation and sequencing

DNA was extracted from frozen samples stored in liquid nitrogen using the DNeasy Blood & Tissue Kit (Qiagen, Cat. No. 69506). DNA quantity was assessed with the DNA BR Assay Kit (Thermo Fisher, Cat. No. Q32853) on a Qubit™ fluorometer. Purified high-molecular-weight, RNA-free DNA (~ 200 ng) was used for whole-exome library preparation with the SureSelect XT HS2 DNA Kit (Agilent). Libraries were validated on the Agilent 2100 Bioanalyzer and sequenced (2 × 150 bp, ~ 100× coverage) on the NovaSeq 6000 (Illumina). Raw FASTQ files were processed using the DRAGEN v4.3 pipeline against GRCh38 (hg38).

Variant annotation, filtering, and clinical classification

Germline VCF files from 479 samples were annotated in VarSeq (Golden Helix, Inc., Bozeman, MT). Variants were filtered for genotype quality ≥ 20, read depth ≥ 10, and minor allele frequency < 5% in population databases; benign/likely benign ClinVar entries were excluded. The filtered variants were further assessed and classified by VarSeq according to guidelines given by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP)¹³. Variants classified as pathogenic or likely pathogenic through VarSeq annotation, or in the ClinVar database, were retained for further analysis.

We analyzed pathogenic and likely pathogenic variants across 97 cancer susceptibility genes, 15 homologous recombination repair (HRR) genes (ATM, BARD1, BRCA1, BRCA2, BRIP1, CDK12, CHEK1, CHEK2, FANCL, PALB2, PPP2R2A, RAD51B, RAD51C, RAD51D, and RAD54L), and 82 other genes linked to hereditary cancer syndromes [12].

To assess clinically actionable pharmacogenomic markers, variants were compared against PharmGKB Level 1 A/1B entries relevant to cancer treatment [13]. Genotypic data from 479 individuals were then categorized by functional impact according to Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines [14].

Population-specific variation in the prevalence of pathogenic variants

To assess population-specific differences in the prevalence of pathogenic variants, we compared allele frequencies observed in the cohort of 479 individuals with those reported in diverse subpopulations from the Genome Aggregation Database (gnomAD v4.1.0)¹⁷. Allele frequencies for variants classified as pathogenic or likely pathogenic identified in the dataset were calculated. Corresponding allele frequencies for the same variants were extracted from gnomAD. The subpopulations analyzed included African (AFR), Latino/Admixed American (AMR), Ashkenazi Jewish (ASJ), East Asian (EAS), Finnish (FIN), Middle Eastern (MID), Non-Finnish European (NFE), South Asian (SAS), and Other (OTH) cohorts.

Assessment of germline RECQL mutations and secondary findings

To assess RECQL as a potential breast cancer susceptibility gene, we curated pathogenic/likely pathogenic variants from ClinVar and prior breast cancer studies, and evaluated their presence in the Indian cohort with allele frequencies compared to gnomAD. Secondary findings were identified by screening for pathogenic/likely pathogenic variants in non-oncology genes from the ACMG SF v3.2 list [15], retaining those classified as pathogenic/likely pathogenic in VarSeq for analysis.

Results

Sequencing metrics summary

Across the 479 samples, sequencing demonstrated high data quality and depth. On average, 87,946,410 raw reads were generated per sample, with 73,854,690 passing quality control. Alignment yielded 86,632,490 mapped reads on average, of which 64,257,120 were unique after duplicate removal, corresponding to a mean duplication rate of 25.7%. Furthermore, an average of 82,852,300 reads had a mapping quality (MAPQ) above 40, achieving a mean on-target coverage depth of 108.2x across all samples. A summary of the sequencing and alignment metrics for the generated data is given in supplementary table 1.

Prevalence of germline variants in cancer susceptibility genes

Filtered variants across the 479 samples were analyzed in this study to identify pathogenic and likely pathogenic variants in genes that are known to be associated with hereditary cancer syndromes or germline predisposition to cancer. After filtering out variants that did not match the expected zygosity and autosomal dominant inheritance pattern, pathogenic and likely pathogenic variants retained were those classified either by VarSeq or ClinVar.

A total of 33 unique pathogenic or likely pathogenic variants were identified across the BRCA1 and BRCA2 genes (Table 1). Among the analyzed cohort, 40 patients carried at least one of these variants, resulting in an overall germline positivity rate for pathogenic or likely pathogenic BRCA1/2 variants in this cohort of 8.35%. In all 40 patients, these variants were heterozygous, consistent with the autosomal dominant inheritance pattern observed in hereditary breast and ovarian cancer syndrome. Notably, certain variants exhibited a relatively high frequency in the cohort. The most recurrent variant was BRCA1:NM_007294.4:c.68_69delAG (p.Glu23Valfs*17), observed in 5 individuals (1.04% of the cohort), which was the most frequently detected variant among these two genes in this cohort.

Table 1.

Pathogenic and likely pathogenic variants identified across 97 cancer predisposition genes in 479 patients

Variant	Gene	Classification	Disorder	ACMG/AMP Attributes (Golden Helix)	Inheritance	Classification Source	Effect (Combined)	HGVS c.	HGVS p.	Total Samples in Cohort	% of Samples in Cohort	Homozygous Samples in Cohort	Heterozygous Samples in Cohort
chr1-161328481-C-T	SDHC	Likely pathogenic	Pheochromocytoma/paraganglioma syndrome 3	Attributes: PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	Missense	NM_003001.5:c.163 C > T	NP_002992.1:p.His55Tyr	2	0.42	0	2
chr2-47466656-A-C	MSH2	Pathogenic	Colorectal cancer hereditary nonpolyposis type 1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000251.3:c.1511–2 A > C		1	0.21	0	1
chr2-47466797-AT-A	MSH2	Pathogenic	Hereditary_nonpolyposis_colorectal_neoplasms	-	Autosomal Dominant	ClinVar	LoF	NM_000251.3:c.1653delT	NP_000242.1:p.Phe551Leufs*6	1	0.21	0	1
chr2-47476378-G-T	MSH2	Pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_000251.3:c.2017G > T	NP_000242.1:p.Gly673Ter	1	0.21	0	1
chr2-47799596-ATCT-A	MSH6	Likely pathogenic	Colorectal cancer hereditary nonpolyposis type 5	Attributes: PM2, PM4, PP3, PP5, PM5	Autosomal Dominant	Golden Helix	Missense	NM_000179.3:c.1618_1620delCTT	NP_000170.1:p.Leu540del	1	0.21	0	1
chr2-47806651-G-T	MSH6	Likely_pathogenic	Lynch_syndrome_5	-	Autosomal Dominant	ClinVar	Missense	NM_000179.3:c.4001G > T	NP_000170.1:p.Arg1334Leu	1	0.21	0	1
chr2-214745855-C-A	BARD1	Pathogenic/Likely_pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_000465.4:c.1678-1G > T		1	0.21	0	1
chr2-214769255-C-CT	BARD1	Pathogenic	Familial_cancer_of_breast	-	Autosomal Dominant	ClinVar	LoF	NM_000465.4:c.1371dupA	NP_000456.2:p.Asp458Argfs*10	1	0.21	0	1
chr3-12584539-G-A	RAF1	Likely pathogenic	Noonan syndrome 5	Attributes: PM2, PP2, PP3, PS1	Autosomal Dominant	Golden Helix	Missense	NM_002880.4:c.1922 C > T	NP_002871.1:p.Thr641Met	1	0.21	0	1
chr3-36993650-A-G	MLH1	Pathogenic	Colorectal cancer hereditary nonpolyposis type 2	Attributes: PM2, PM1, PP2, PP3, PS1, PM5	Autosomal Dominant	Golden Helix	Missense	NM_000249.4:c.103 A > G	NP_000240.1:p.Met35Val	1	0.21	0	1
chr3-36996626-G-T	MLH1	Likely pathogenic	Colorectal cancer hereditary nonpolyposis type 2	Attributes: PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	Missense	NM_000249.4:c.124G > T	NP_000240.1:p.Ala42Ser	1	0.21	0	1
chr3-37001037-A-G	MLH1	Likely pathogenic	Colorectal cancer hereditary nonpolyposis type 2	Attributes: PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	LoF	NM_000249.4:c.290 A > G	NP_000240.1:p.Tyr97Cys	7	1.46	0	7
chr3-37008813-G-T	MLH1	Likely_pathogenic	Lynch_syndrome|Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_000249.4:c.454-1G > T		1	0.21	0	1
chr3-37014544-C-T	MLH1	Likely pathogenic	Colorectal cancer hereditary nonpolyposis type 2	Attributes: PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	Missense	NM_000249.4:c.790 C > T	NP_000240.1:p.His264Tyr	1	0.21	0	1
chr3-37025635-A-T	MLH1	Likely_pathogenic	Hereditary_cancer-predisposing_syndrome|Lynch_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_000249.4:c.1039–2 A > T		1	0.21	0	1
chr3-37047557-A-C	MLH1	Likely pathogenic	Colorectal cancer hereditary nonpolyposis type 2	Attributes: PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	Missense	NM_000249.4:c.1770 A > C	NP_000240.1:p.Leu590Phe	1	0.21	0	1
chr3-37048952-T-G	MLH1	Likely pathogenic	Colorectal cancer hereditary nonpolyposis type 2	Attributes: PM2, PM1, PP2, PP3, PM5	Autosomal Dominant	Golden Helix	Missense	NM_000249.4:c.2038T > G	NP_000240.1:p.Cys680Gly	4	0.84	0	4
chr3-52408603-AGGGCTG-A	BAP1	Likely pathogenic	Tumor predisposition syndrome 1	Attributes: PM2, PVS1	Autosomal Dominant	Golden Helix	LoF	NM_004656.4:c.123-3_125delCAGCCC	NP_004647.1:p.?	1	0.21	0	1
chr5-228183-A-T	SDHA	Likely_pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_004168.4:c.622–2 A > T		1	0.21	0	1
chr5-236581-G-T	SDHA	Pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_004168.4:c.1414G > T	NP_004159.2:p.Glu472Ter	1	0.21	0	1
chr5-254392-G-T	SDHA	Pathogenic/Likely_pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_004168.4:c.1795-1G > T		1	0.21	0	1
chr5-112767391-GTA-G	APC	Likely_pathogenic	Familial_adenomatous_polyposis_1	-	Autosomal Dominant	ClinVar	LoF	NM_000038.6:c.422 + 2_422 + 3delTA		2	0.42	0	2
chr5-112827175-CT-C	APC	Pathogenic	Familial_adenomatous_polyposis_1	-	Autosomal Dominant	ClinVar	LoF	NM_000038.6:c.1477delT	NP_000029.2:p.Tyr493Thrfs*5	1	0.21	0	1
chr6-35452658-AG-A	FANCE	Pathogenic/Likely_pathogenic	Fanconi_anemia_complementation_group_E	-	Autosomal Recessive	ClinVar	LoF	NM_021922.3:c.118delG	NP_068741.1:p.Ala40Argfs*44	1	0.21	1	0
chr7-6009009-G-C	PMS2	Likely pathogenic	Lynch syndrome 4	Attributes: PM2, PS1	Autosomal Dominant	Golden Helix	Missense	NM_000535.7:c.11 C > G	NP_000526.2:p.Ala4Gly	1	0.21	0	1
chr7-140834827-CTCTTTGTTGG-C	BRAF	Likely pathogenic	Noonan syndrome 7	Attributes: PM2, PVS1	Autosomal Dominant	Golden Helix	LoF	NM_004333.6:c.276_285delCCAACAAAGA	NP_004324.2:p.Gln93Asnfs*86	1	0.21	0	1
chr9-21970951-G-GC	CDKN2A	Likely pathogenic	Melanoma-pancreatic cancer syndrome	Attributes: PM2, PVS1_Strong	Autosomal Dominant	Golden Helix	LoF	NM_000077.5:c.407dupG	NP_000068.1:p.Thr137Hisfs*5	1	0.21	0	1
chr10-43100718-CCG-C	RET	Likely pathogenic	Retinoblastoma	Attributes: PM2, PVS1	Autosomal Dominant	Golden Helix	LoF	NM_020975.6:c.335_336delGC	NP_066124.1:p.Arg112Glnfs*22	1	0.21	0	1
chr10-86890064-C-T	BMPR1A	Pathogenic	Juvenile_polyposis_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_004329.3:c.70 C > T	NP_004320.2:p.Gln24Ter	1	0.21	0	1
chr10-86919324-G-T	BMPR1A	Likely_pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	Missense	NM_004329.3:c.1021G > T	NP_004320.2:p.Gly341Cys	1	0.21	0	1
chr10-86919360-C-T	BMPR1A	Pathogenic	Juvenile_polyposis_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_004329.3:c.1057 C > T	NP_004320.2:p.Gln353Ter	1	0.21	0	1
chr10-87952234-TC-T	PTEN	Pathogenic	Cowden_syndrome_1	-	Autosomal Dominant	ClinVar	LoF	NM_000314.8:c.611delC	NP_000305.3:p.Pro204Glnfs*17	1	0.21	0	1
chr10-87960893-G-T	PTEN	Pathogenic	Cowden_syndrome_1	-	Autosomal Dominant	ClinVar	LoF	NM_000314.8:c.802-1G > T		1	0.21	0	1
chr11-532524-T-G	HRAS	Likely pathogenic	Costello syndrome	Attributes: PM2, PVS1, BP6	Autosomal Dominant	Golden Helix	LoF	NM_005343.4:c.*6–2 A > C		1	0.21	0	1
chr11-64805126-T-C	MEN1	Likely pathogenic	Multiple endocrine neoplasia 1	Attributes: PM2, PM1, PP2, PP3, PM5	Autosomal Dominant	Golden Helix	Missense	NM_130799.3:c.1258 A > G	NP_570711.2:p.Ile420Val	1	0.21	0	1
chr11-64806286-C-T	MEN1	Likely pathogenic	Multiple endocrine neoplasia 1	Attributes: PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	Missense	NM_130799.3:c.995G > A	NP_570711.2:p.Arg332His	1	0.21	0	1
chr11-108289762-G-C	ATM	Pathogenic	Breast cancer susceptibility to	Attributes: PM2, PM1, PP2, PP3, PS1	Autosomal Dominant	Golden Helix	Missense	NM_000051.4:c.4397G > C	NP_000042.3:p.Arg1466Pro	1	0.21	0	1
chr11-108294927-G-T	ATM	Pathogenic	Familial_cancer_of_breast	-	Autosomal Dominant	ClinVar	LoF	NM_000051.4:c.4777G > T	NP_000042.3:p.Glu1593Ter	1	0.21	0	1
chr11-108299873-TG-T	ATM	Pathogenic	Familial_cancer_of_breast	-	Autosomal Dominant	ClinVar	LoF	NM_000051.4:c.5167delG	NP_000042.3:p.Val1723Ter	1	0.21	0	1
chr11-108301790-G-A	ATM	Likely_pathogenic	Familial_cancer_of_breast	-	Autosomal Dominant	ClinVar	LoF	NM_000051.4:c.5319 + 1G > A		1	0.21	0	1
chr11-108316016-G-A	ATM	Likely pathogenic	Breast cancer susceptibility to	Attributes: BS2, PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	Missense	NM_000051.4:c.6101G > A	NP_000042.3:p.Arg2034Gln,	1	0.21	0	1
chr11-108326124-C-T	ATM	Pathogenic	Familial_cancer_of_breast	-	Autosomal Dominant	ClinVar	LoF	NM_000051.4:c.6874 C > T	NP_000042.3:p.Gln2292Ter,	1	0.21	0	1
chr11-108329198-G-T	ATM	Pathogenic	Familial_cancer_of_breast	-	Autosomal Dominant	ClinVar	LoF	NM_000051.4:c.7267G > T	NP_000042.3:p.Glu2423Ter,	1	0.21	0	1
chr11-119278287-C-T	CBL	Likely pathogenic	Noonan syndrome-like disorder with or without juvenile myelomonocytic leukemia	Attributes: PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	Missense	NM_005188.4:c.1217 C > T	NP_005179.2:p.Thr406Ile	1	0.21	0	1
chr11-119278514-C-T	CBL	Likely pathogenic	Noonan syndrome-like disorder with or without juvenile myelomonocytic leukemia	Attributes: PM2, PM1, PP2, PP3	Autosomal Dominant	Golden Helix	Missense	NM_005188.4:c.1232 C > T	NP_005179.2:p.Ser411Leu	1	0.21	0	1
chr12-25245350-C-A	KRAS	Pathogenic	Noonan syndrome 3	-	Autosomal Dominant	ClinVar	Missense	NM_004985.5:c.35G > T	NP_004976.2:p.Gly12Val	1	0.21	0	1
chr13-32316508-GAC-G	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.51_52delAC	NP_000050.3:p.Arg18Leufs*12	1	0.21	0	1
chr13-32333010-C-A	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.1532 C > A	NP_000050.3:p.Ser511Ter	1	0.21	0	1
chr13-32333292-T-TA	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000059.4:c.1815dupA	NP_000050.3:p.Pro606Thrfs*10	1	0.21	0	1
chr13-32336794-TC-T	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.2442delC	NP_000050.3:p.Met815Trpfs*10	1	0.21	0	1
chr13-32337273-C-A	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.2918 C > A	NP_000050.3:p.Ser973Ter	2	0.42	0	2
chr13-32337570-T-G	BRCA2	Pathogenic	Breast-ovarian cancer familial 2	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000059.4:c.3215T > G	NP_000050.3:p.Leu1072Ter	1	0.21	0	1
chr13-32337870-C-A	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.3515 C > A	NP_000050.3:p.Ser1172Ter	1	0.21	0	1
chr13-32337996-TG-T	BRCA2	Likely pathogenic	Breast-ovarian cancer familial 2	Attributes: PM2, PVS1	Autosomal Dominant	Golden Helix	LoF	NM_000059.4:c.3645delG	NP_000050.3:p.Phe1216Leufs*12	1	0.21	0	1
chr13-32338058-C-T	BRCA2	Pathogenic/Likely_pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.3703 C > T	NP_000050.3:p.Gln1235Ter	1	0.21	0	1
chr13-32338167-C-G	BRCA2	Pathogenic	Breast-ovarian cancer familial 2	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000059.4:c.3812 C > G	NP_000050.3:p.Ser1271Ter	1	0.21	0	1
chr13-32338886-G-T	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.4531G > T	NP_000050.3:p.Glu1511Ter	1	0.21	0	1
chr13-32339924-G-T	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.5569G > T	NP_000050.3:p.Glu1857Ter	1	0.21	0	1
chr13-32339932-T-TA	BRCA2	Pathogenic	Breast-ovarian cancer familial 2	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000059.4:c.5583dupA	NP_000050.3:p.Val1862Serfs*11	2	0.42	0	2
chr13-32340089-G-T	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.5734G > T	NP_000050.3:p.Glu1912Ter	1	0.21	0	1
chr13-32340419-TCAGAC-T	BRCA2	Pathogenic	Breast-ovarian cancer familial 2	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000059.4:c.6068_6072delACCAG	NP_000050.3:p.Asp2023Alafs*24	1	0.21	0	1
chr13-32362695-C-A	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.7976 + 2 C > A		1	0.21	0	1
chr13-32379503-GA-G	BRCA2	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _2	-	Autosomal Dominant	ClinVar	LoF	NM_000059.4:c.8946delA	NP_000050.3:p.Asp2983Ilefs*5	1	0.21	0	1
chr13-32394685-CTAGGA-C	BRCA2	Pathogenic	Breast-ovarian cancer familial 2	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000059.4:c.9257-3_9258delTAGGA	NP_000050.3:p.?	1	0.21	0	1
chr13-48412354-A-ACATG	RB1	Pathogenic	Retinoblastoma	-	Autosomal Dominant	ClinVar	LoF	NM_001162498.3:c.66_69dupCATG	NP_001155970.1:p.Phe24Hisfs*29,	5	1.04	0	5
chr13-48459792-C-T	RB1	Pathogenic	Retinoblastoma	-	Autosomal Dominant	ClinVar	LoF	NM_000321.3:c.2065 C > T	NP_000312.2:p.Gln689Ter	1	0.21	0	1
chr13-48464984-C-G	RB1	Pathogenic	Retinoblastoma	-	Autosomal Dominant	ClinVar	Other	NM_000321.3:c.2212–14 C > G		1	0.21	0	1
chr14-95124428-C-A	DICER1	Pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_177438.3:c.1144G > T	NP_803187.1:p.Glu382Ter	2	0.42	0	2
chr16-23603500-C-A	PALB2	Likely pathogenic	Breast-ovarian cancer, familial, susceptibility to, 5	Attributes: PM2, PVS1_Strong	Autosomal Dominant	Golden Helix	LoF	NM_024675.4:c.3520G > T	NP_078951.2:p.Gly1174Ter	1	0.21	0	1
chr16-23629665-TC-T	PALB2	Pathogenic	Breast-ovarian cancer, familial, susceptibility to, 5	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_024675.4:c.2488delG	NP_078951.2:p.Glu830Serfs*21	1	0.21	0	1
chr16-23629980-G-C	PALB2	Pathogenic	Breast-ovarian cancer, familial, susceptibility to, 5	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_024675.4:c.2174 C > G	NP_078951.2:p.Ser725Ter	1	0.21	0	1
chr16-23635536-A-T	PALB2	Pathogenic	Breast-ovarian cancer, familial, susceptibility to, 5	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_024675.4:c.1010T > A	NP_078951.2:p.Leu337Ter	1	0.21	0	1
chr16-68810195-A-G	CDH1	Pathogenic	Gastric_cancer	-	Autosomal Dominant	ClinVar	LoF	NM_004360.5:c.688–2 A > G		1	0.21	0	1
chr17-7673788-G-A	TP53	Pathogenic/Likely_pathogenic	Li-Fraumeni_syndrome	-	Autosomal Dominant	ClinVar	Missense	NM_000546.6:c.832 C > T	NP_000537.3:p.Pro278Ser	1	0.21	0	1
chr17-7673803-G-A	TP53	Pathogenic/Likely_pathogenic	Li-Fraumeni_syndrome	-	Autosomal Dominant	ClinVar	Missense	NM_000546.6:c.817 C > T	NP_000537.3:p.Arg273Cys	1	0.21	0	1
chr17-7674947-A-G	TP53	Pathogenic	Li-Fraumeni_syndrome	-	Autosomal Dominant	ClinVar	Missense	NM_000546.6:c.584T > C	NP_000537.3:p.Ile195Thr	1	0.21	0	1
chr17-7675070-C-T	TP53	Likely pathogenic	Li-Fraumeni_syndrome	Attributes: PM2, PP2, PS1, PM5	Autosomal Dominant	Golden Helix	Missense	NM_000546.6:c.542G > A	NP_000537.3:p.Arg181His	1	0.21	0	1
chr17-7675139-C-A	TP53	Pathogenic/Likely_pathogenic	Li-Fraumeni_syndrome	-	Autosomal Dominant	ClinVar	Missense	NM_000546.6:c.473G > T	NP_000537.3:p.Arg158Leu	1	0.21	0	1
chr17-7675217-T-C	TP53	Pathogenic/Likely_pathogenic	Li-Fraumeni_syndrome	-	Autosomal Dominant	ClinVar	Missense	NM_000546.6:c.395 A > G	NP_000537.3:p.Lys132Arg	1	0.21	0	1
chr17-7675237-C-A	TP53	Pathogenic	Li-Fraumeni_syndrome	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000546.6:c.376-1G > T		1	0.21	0	1
chr17-31095370-G-A	NF1	Pathogenic	Neurofibromatosis, _type_1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000267.3:c.60 + 1G > A		1	0.21	0	1
chr17-31169898-G-T	NF1	Pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	LoF	NM_000267.3:c.487G > T	NP_000258.1:p.Glu163Ter	1	0.21	0	1
chr17-31181717-G-A	NF1	Pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	LoF	NM_000267.3:c.662G > A	NP_000258.1:p.Trp221Ter	1	0.21	0	1
chr17-31226456-G-T	NF1	Pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	LoF	NM_000267.3:c.2023G > T	NP_000258.1:p.Gly675Ter	1	0.21	0	1
chr17-31233116-G-T	NF1	Pathogenic/Likely_pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	Missense	NM_000267.3:c.3611G > T	NP_000258.1:p.Arg1204Leu	1	0.21	0	1
chr17-31248982-A-G	NF1	Pathogenic	Neurofibromatosis, _type_1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000267.3:c.3975–2 A > G		1	0.21	0	1
chr17-31265242-G-T	NF1	Pathogenic/Likely_pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	LoF	NM_000267.3:c.4675G > T	NP_000258.1:p.Glu1559Ter	1	0.21	0	1
chr17-31265336-G-T	NF1	Pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	Missense	NM_000267.3:c.4769G > T	NP_000258.1:p.Arg1590Leu	1	0.21	0	1
chr17-31327839-G-T	NF1	Pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	Missense	NM_000267.3:c.5546G > T	NP_000258.1:p.Arg1849Leu	1	0.21	0	1
chr17-31330351-G-T	NF1	Pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	LoF	NM_000267.3:c.5602G > T	NP_000258.1:p.Glu1868Ter	1	0.21	0	1
chr17-31330456-TTAGAATTTTTGGAAG-T	NF1	Likely pathogenic	Neurofibromatosis, _type_1	Attributes: PM2, PVS1	Autosomal Dominant	Golden Helix	LoF	NM_000267.3:c.5708_5722delTAGAATTTTTGGAAG	NP_000258.1:p.Leu1903Ter	1	0.21	0	1
chr17-31334843-G-T	NF1	Pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	LoF	NM_000267.3:c.5755G > T	NP_000258.1:p.Glu1919Ter	1	0.21	0	1
chr17-31358481-C-T	NF1	Likely_pathogenic	Neurofibromatosis, _type_1	-	Autosomal Dominant	ClinVar	Missense	NM_000267.3:c.7909 C > T	NP_000258.1:p.His2637Tyr	1	0.21	0	1
chr17-43045767-G-A	BRCA1	Likely pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1_Strong, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.5503 C > T	NP_009225.1:p.Arg1835Ter	2	0.42	0	2
chr17-43045803-C-T	BRCA1	Likely pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1_Strong, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.5468-1G > A		1	0.21	0	1
chr17-43063930-C-T	BRCA1	Pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PP2, PP3, PS1, PM5	Autosomal Dominant	Golden Helix	Missense	NM_007294.4:c.5096G > A	NP_009225.1:p.Arg1699Gln	1	0.21	0	1
chr17-43067646-AG-A	BRCA1	Pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.5035delC	NP_009225.1:p.Leu1679Ter	1	0.21	0	1
chr17-43076487-C-T	BRCA1	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _1	-	Autosomal Dominant	ClinVar	LoF	NM_007294.4:c.4484 + 1G > A		1	0.21	0	1
chr17-43090962-ACT-A	BRCA1	Pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.4165_4166delAG	NP_009225.1:p.Ser1389Ter	1	0.21	0	1
chr17-43092050-CCTTT-C	BRCA1	Pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.3477_3480delAAAG	NP_009225.1:p.Ile1159Metfs*50	1	0.21	0	1
chr17-43092698-TAC-T	BRCA1	Likely pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.2831_2832delGT	NP_009225.1:p.Cys944Ter	1	0.21	0	1
chr17-43092792-ATTCT-A	BRCA1	Pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.2735_2738delAGAA	NP_009225.1:p.Lys912Metfs*87	1	0.21	0	1
chr17-43093742-C-A	BRCA1	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _1	-	Autosomal Dominant	ClinVar	LoF	NM_007294.4:c.1789G > T	NP_009225.1:p.Glu597Ter	1	0.21	0	1
chr17-43094342-CA-C	BRCA1	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _1	-	Autosomal Dominant	ClinVar	LoF	NM_007294.4:c.1188delT	NP_009225.1:p.Asp396Glufs*14	1	0.21	0	1
chr17-43094381-C-A	BRCA1	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _1	-	Autosomal Dominant	ClinVar	LoF	NM_007294.4:c.1150G > T	NP_009225.1:p.Glu384Ter	1	0.21	0	1
chr17-43094522-C-A	BRCA1	Pathogenic	Breast-ovarian_cancer, _familial, _susceptibility_to, _1	-	Autosomal Dominant	ClinVar	LoF	NM_007294.4:c.1009G > T	NP_009225.1:p.Glu337Ter	1	0.21	0	1
chr17-43106528-C-CA	BRCA1	Pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.139dupT	NP_009225.1:p.Cys47Leufs*19	1	0.21	0	1
chr17-43124027-ACT-A	BRCA1	Pathogenic	Breast-ovarian cancer familial 1	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007294.4:c.68_69delAG	NP_009225.1:p.Glu23Valfs*17	5	1.04	0	5
chr18-51067167-T-TA	SMAD4	Pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	LoF	NM_005359.6:c.1289dupA	NP_005350.1:p.Tyr430Ter	1	0.21	0	1
chr19-1207163-A-T	STK11	Pathogenic	Peutz-Jeghers syndrome	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_000455.5:c.250 A > T	NP_000446.1:p.Lys84Ter	1	0.21	0	1
chr22-23787210-C-A	SMARCB1	Pathogenic/Likely_pathogenic	Hereditary_cancer-predisposing_syndrome	-	Autosomal Dominant	ClinVar	Missense	NM_003073.5:c.41 C > A	NP_003064.2:p.Pro14His	1	0.21	0	1
chr22-28695175-C-A	CHEK2	Pathogenic	Familial_cancer_of_breast	-	Autosomal Dominant	ClinVar	LoF	NM_007194.4:c.1327G > T	NP_009125.1:p.Gly443Ter	1	0.21	0	1
chr22-28711914-C-A	CHEK2	Pathogenic	Familial_cancer_of_breast	Attributes: PM2, PVS1, PP5	Autosomal Dominant	Golden Helix	LoF	NM_007194.4:c.787G > T	NP_009125.1:p.Glu263Ter	1	0.21	0	1
chr22-28711986-C-T	CHEK2	Likely pathogenic	Familial_cancer_of_breast	Attributes: PM2, PM1, PS1	Autosomal Dominant	Golden Helix	Missense	NM_007194.4:c.715G > A	NP_009125.1:p.Glu239Lys	1	0.21	0	1
chr22-28725089-T-C	CHEK2	Likely pathogenic	Familial_cancer_of_breast	Attributes: BS2, PM2, PM1, PM5	Autosomal Dominant	Golden Helix	Missense	NM_007194.4:c.480 A > G	NP_009125.1:p.Ile160Met	1	0.21	0	1

Among the 15 genes in the Homologous Recombination Repair (HRR) pathway, pathogenic and likely pathogenic variants were identified in a total of 6 genes (BARD1, ATM, BRCA2, PALB2, BRCA1, and CHEK2). A total of 50 unique variants were detected across 57 patients, representing a germline positivity rate of 11.9% for HRR genes (Table 1). In terms of distribution, BRCA2 variants were the most frequent (36% of the 50 unique variants), followed by BRCA1 (30%) and ATM (14%). No variants were detected in BRIP1, CDK12, CHEK1, FANCL, PPP2R2A, RAD51B, RAD51C, RAD51D, and RAD54L.

Additionally, pathogenic and likely pathogenic variants were identified in a total of 33 genes associated with hereditary cancer syndromes or predisposition, considering the broader panel of 97 genes (Table 1). Across the cohort, 115 unique variants were detected in 118 patients, leading to an overall germline positivity of 24.6% across this panel. Of these, 102 patients carried only one pathogenic or likely pathogenic variant, while 14 individuals had two such variants, and 3 patients carried three variants. Apart from HRR genes, the most frequently affected genes included MLH1 (3.5%), NF1 (2.7%), TP53 (1.5%), and RB1 (1.5%). The remaining genes collectively accounted for 7.7% of the cohort. The distribution and types of pathogenic or likely pathogenic (P/LP) variant effects per gene across all samples is summarized in Fig. 1A and B.

Fig. 1.

Distribution of Pathogenic/Likely Pathogenic Variants Across Genes in Primary and Incidental Findings. (A) Stacked bar plot showing the number and type of pathogenic or likely pathogenic (P/LP) variant effects per gene for all samples. (B) Donut chart depicting percentage distribution of P/LP variants among cancer susceptibility genes and (C) non-oncological incidental findings related genes

Genetic landscape of pathogenic variants

A comparative analysis of allele frequencies was conducted between the cohort of 479 individuals and various populations represented in the gnomAD database (Supplementary table 2).

Among the pathogenic and likely pathogenic variants identified in the HRR pathway genes cohort, 31 variants were completely absent in all gnomAD subpopulations, suggesting the presence of potentially novel or rare disease-associated alleles in this cohort (Fig. 2A). Two variants, BRCA2:NM_000059.4:c.2918 C > A, and BRCA2:NM_000059.4:c.5583dupA, were detected at a measurable frequency (0.21%) in the Indian cohort but were absent in gnomAD subpopulations, indicating their rarity in broader global populations. The variant BRCA1:NM_007294.4:c.5503 C > T has an allele frequency of 0.21% in the cohort, while its frequency in other populations is less than 0.003%.

Fig. 2.

Plot showing the frequency distribution of pathogenic or likely pathogenic variants across different gnomAD populations and WES cohort data in (A) genes related to cancer predisposition, and (B) in all remaining genes related to cancer predisposition. Variants that were classified as pathogenic/likely pathogenic by only ClinVar are labeled in red gnomAD - AFR: African/African American, AMR: Latino/Admixed American, ASJ: Ashkenazi Jewish, EAS: East Asian, FIN: Finnish, MID: Middle Eastern, NFE: Non-Finnish European, OTH: Other, SAS: South Asian

Conversely, BRCA1:NM_007294.4:c.68_69delAG, a well-established pathogenic variant [16], was observed at an allele frequency of 0.52% in the Indian cohort, which is significantly higher than its reported frequencies in other gnomAD populations. In the gnomAD Ashkenazi Jewish (ASJ) subpopulation, this variant was also detected at a notably elevated frequency of 0.42%, which is consistent with previous reports indicating an increased prevalence of BRCA1 founder mutations in individuals of Ashkenazi Jewish ancestry.

Among the pathogenic and likely pathogenic variants identified in the remaining genes, 41 were seen to be completely absent in all gnomAD subpopulations but were detected at a measurable frequency in the Indian cohort, suggesting their rarity in global datasets (Fig. 2B). The variants APC:NM_000038.6:c.422 + 2_422 + 3delTA, DICER1:NM_177438.3:c.1144G > T, FANCE:NM_021922.3:c.118delG, and SDHC:NM_003001.5:c.163 C > T have an allele frequency of 0.21% in the Indian cohort, while they are absent in other populations.

Several variants showed notable frequency differences. MLH1:NM_000249.4:c.2038T > G (0.42% vs. 0.0032% in gnomAD; highest SAS 0.054%) and MLH1:NM_000249.4:c.290 A > G (0.73% vs. 0.011% overall; highest SAS 0.18%) were enriched. RB1:NM_001162498.3:c.66_69dupCATG (0.52% vs. 0.0021% overall; SAS only) also showed enrichment. Conversely, RB1:NM_000321.3:c.2212–14 C > G was lower in this cohort (0.10%) compared to NFE (0.56%), SAS (0.15%), ASJ (0.16%), and AMR (0.086%).

Germline RECQL variants

A total of 69 unique RECQL gene variants, curated from different publications and ClinVar (Supplementary table 3), were evaluated for their prevalence in our cohort of unselected breast cancer patients. Of these, 2 variants, NM_032941.3:c.644G > A and NM_032941.3:c.546 C > T, were identified in the Indian cohort in a total of 8 individuals, translating to a positivity rate of 1.7%. The remaining 67 variants were not observed in the Indian cohort. Comparison of allele frequencies of these variants with population data from gnomAD suggested that a substantial proportion of RECQL variants reported in the literature also occur at low frequencies in general populations (Fig. 3, supplementary table 4).

Fig. 3.

Prevalence of germline RECQL mutations in different populations gnomAD - AFR: African/African American, AMR: Latino/Admixed American, ASJ: Ashkenazi Jewish, EAS: East Asian, FIN: Finnish, MID: Middle Eastern, NFE: Non-Finnish European, OTH: Other, SAS: South Asian

Of the two RECQL variants identified, the c.644G >A variant was observed in 7 individuals, corresponding to an allele frequency of 0.73% in the Indian breast cancer cohort. In comparison, its frequency in the gnomAD database is considerably lower, with an overall population frequency of 0.01%, and notably higher representation in the South Asian subpopulation (0.17%), suggesting possible enrichment in individuals of South Asian ancestry. The second variant, c.546 C >T, was detected in 1 individual (allele frequency 0.1%) and is absent in all gnomAD subpopulations, including South Asians. Previously reported population-enriched RECQL variants (c.643 C >T in Quebec, c.1667_1667 + 3delAGTA in Poland, and c.1088 A >G, c.2T >C, c.199G >A in China) [17, 18].

Non-oncologic germline incidental findings

To assess the prevalence of incidental or secondary findings, we screened for pathogenic and likely pathogenic variants in non-oncology genes listed in the ACMG SF v3.2 recommendations. In our cohort, we identified 64 unique pathogenic/likely pathogenic variants across 27 genes, affecting a total of 104 individuals. These variants were associated with 20 distinct inherited disorders, many of which have established guidelines for clinical surveillance or intervention. Notably, 9 individuals carried two distinct actionable variants, underscoring the potential for multiple secondary findings in a single genome. Of the variants detected, 55 were linked to 16 autosomal dominant conditions, affecting 66 patients, while 40 individuals were identified as carriers for three autosomal recessive disorders. This corresponds to an overall secondary findings positivity rate of 21.7% (104/479 individuals), with 13.8% carrying dominant variants of potential personal health significance and 8.4% identified as carriers.

Among the dominant conditions, the most frequently implicated gene was FBN1, associated with Marfan syndrome, found in 16 individuals. RYR1, linked to malignant hyperthermia susceptibility, was identified in 9 individuals, and SCN5A, associated with cardiac arrhythmias such as long QT syndrome type 3, Brugada syndrome, and dilated cardiomyopathy, was seen in 5 cases. LDLR mutations related to familial hypercholesterolemia were also seen in 5 patients.

Among recessive carrier states, BTD, the gene associated with biotinidase deficiency, was the most common, with 30 individuals identified as carriers. ATP7B, the gene linked to Wilson disease, was found in a heterozygous state in 7 individuals. Detailed information on the variants, associated conditions, and inheritance patterns is provided in supplementary table 5. The percentage distribution of P/LP variants among genes associated with non-oncological incidental findings is summarized in Fig. 1C.

Analysis of germline pharmacogenomic variants

Two clinically actionable DPYD variants associated with fluoropyrimidine-related adverse drug reactions were identified in the cohort (Table 2). The DPYD variant rs56038477 (c.1679T > G) was identified in 15 patients (3.13%) with a heterozygous CT genotype. This variant, a marker for the HapB3 allele, is classified as having decreased function, and its presence has been associated with an increased risk of fluoropyrimidine toxicity, particularly in patients receiving capecitabine or fluorouracil.

Table 2.

DPYD Pharmacogenomic variants and their phenotypic implications

Variant	Genotype	Variant ID	Drug(s)	Phenotype Category	Implication	Allele Function	Total Samples in Cohort
chr1:97450058 C/T	CT	rs3918290	capecitabine, fluorouracil, tegafur	Adverse Drug Reaction	Increased Risk	No Function	8
chr1:97573863 C/T	CT	rs56038477	capecitabine, fluorouracil	Adverse Drug Reaction	Increased Risk	Decreased Function	15

Another variant, DPYD rs3918290 (c.1905 + 1G > A) variant was detected in 8 patients (1.67% of 479 patients) in a heterozygous CT genotype. The T allele is classified as ‘no function’ by (CPIC, and individuals carrying this variant may have an increased risk of severe toxicity when treated with fluoropyrimidine-based chemotherapy, such as capecitabine, fluorouracil, and tegafur. However, conflicting evidence exists regarding the extent of risk, and additional genetic and clinical factors may influence toxicity outcomes.

Discussion

In our study, we performed germline multi-gene panel testing on 479 unselected breast cancer patients and identified pathogenic or likely pathogenic variants in 24.6% of individuals. Of the total number of patients, only 8.35% harbored mutations in BRCA1 or BRCA2, while the remainder were distributed across a spectrum of non-BRCA cancer susceptibility genes, particularly those involved in homologous recombination repair (HRR), mismatch repair (MMR), and tumor suppressor pathways.

Our findings are broadly consistent with other Indian studies, although differences in cohort selection impact positivity rates. For instance, Mittal et al. reported 18.6% positivity in an unselected cohort of 236 breast cancer patients, with mutations most commonly found in BRCA1 (46.8%) and BRCA2 (19.1%), and 34% of pathogenic variants occurring in non-BRCA genes [19]. In contrast, studies of more selected populations, such as those by Pramanik et al. (41.4%), Chheda et al. (31.9%), and Chikkala et al. (29.1%), have reported higher positivity rates, reflecting their enrichment for patients with strong family history and early age at onset [20–22]. While our BRCA1/2 positivity rate is modest compared to these referral-based cohorts, it is still significantly higher than that reported in unselected Western populations (typically 5–10%), supporting the notion of a higher inherited burden of disease in Indian breast cancer patients. The relatively high yield of HRR pathway variants (11.9%) underscores the relevance of broad panels that extend beyond BRCA1/2, particularly given the therapeutic implications for PARP inhibitors and platinum-based chemotherapy.

A recurring pattern across Indian datasets, including ours, is the predominance of BRCA1 mutations over BRCA2, and the presence of recurrent variants such as c.68_69delAG possibly reflecting population-specific or founder effects. Importantly, 67% of positive findings in our study were in non-BRCA genes [23]. Notable among these were variants in MLH1 (3.5%), NF1 (2.7%), TP53 (1.5%), and RB1 (1.5%). MLH1 is classically associated with Lynch syndrome, where colorectal and endometrial cancers predominate. Although its role in breast cancer remains uncertain, the enrichment of MLH1 variants in our cohort supports its inclusion in multi-gene testing panels [24]. NF1 mutations confer a ~ 5-fold increased breast cancer risk before age 50 and a 3.5-fold lifetime risk, and our finding of NF1 variants in 2.7% of patients reinforces its relevance in early-onset Indian breast cancers [25–28]. Germline TP53 mutations, while modest in frequency in our cohort, remain clinically significant given their implications for Li-Fraumeni syndrome surveillance [29, 30]. RB1, best known for retinoblastoma predisposition, has also been implicated in sporadic breast carcinomas, and its detection here suggests the need for further study [31, 32].

In addition to well-established hereditary cancer genes, our analysis identified variants in oncogenes (HRAS, KRAS, BRAF) and in genes primarily linked to non-breast cancer syndromes (MEN1, RET). HRAS, KRAS, and BRAF variants are typically somatic drivers rather than germline predisposition alleles, and their detection here may represent rare germline events of uncertain significance. Notably, germline HRAS mutations cause Costello syndrome, which includes tumor predisposition, suggesting a possible though unconventional link to cancer risk. Similarly, while MEN1 classically predisposes to endocrine tumors, population-based data show nearly a 2-fold increased breast cancer risk in women with MEN1 mutations, alongside frequent loss of menin expression in tumors, implicating a potential role in mammary carcinogenesis [33]. In contrast, inactivating variants in RET (associated with Hirschsprung’s disease) currently lack evidence for breast cancer predisposition. These findings highlight the need for cautious interpretation, clinical correlation, and further studies to clarify the role of such germline variants in breast cancer.

Comparison with gnomAD revealed 31 HRR and 41 non-HRR variants absent from global reference datasets but detected in our cohort, suggesting potential population-specific alleles. Classic BRCA1 founder variants showed elevated frequencies similar to those reported in Ashkenazi Jewish populations. It is well-documented that Ashkenazi Jewish individuals have a higher risk of hereditary breast and ovarian cancer, primarily due to the increased frequency of specific BRCA1 and BRCA2 founder mutations [34]. Such population-level enrichments highlight the critical role of ancestry in assessing hereditary cancer risk and designing genetic screening strategies. Accordingly, our findings underscore the need for ancestry-aware variant interpretation frameworks specifically tailored to Indian populations.

The use of comprehensive, unbiased germline testing allowed us to evaluate genes like RECQL, which have emerging but not yet universally established roles in breast cancer susceptibility. Only two variants reported in literature, c.644G >A and c.546 C >T, were identified in a total of 8 individuals (1.7%). The c.644G >A variant showed possible enrichment in South Asians, while c.546 C >T was absent in all gnomAD subpopulations. Notably, RECQL variants recurrently observed in other populations (c.643 C >T in Quebec, c.1667_1667 + 3delAGTA in Poland, and c.1088 A >G, c.2T >C, c.199G >A in China) were not seen in our cohort, supporting the hypothesis of population-specificity [17, 35]. While RECQL’s role as a breast cancer gene remains under investigation, our findings support its prioritization for further study in Indian patients. These findings demonstrate how broad germline testing can also support the investigation of newer or understudied susceptibility genes in diverse populations.

Beyond cancer predisposition, we identified actionable secondary findings in 21.7% of individuals. Dominant variants were most frequent in FBN1, RYR1, SCN5A, and LDLR, while carrier states were most common in BTD and ATP7B. These results demonstrate the additional health benefits of germline sequencing, consistent with ACMG SF v3.2 recommendations. Clinically actionable DPYD variants were also detected (c.1679T >G in 3.1% and c.1905 + 1G >A in 1.7%), highlighting the utility of integrating pharmacogenomic testing into routine germline workflows to optimize chemotherapy safety [36]. DPYD testing is recommended by the United States Food and Drug Administration (FDA), the European Society for Medical Oncology (ESMO), and the 2025 National Comprehensive Cancer Network (NCCN) guidelines, particularly for patients receiving fluoropyrimidine-based chemotherapy in solid tumors, due to the increased risk of severe toxicity associated with these variants. This highlights the utility of integrating pharmacogenomic testing into routine germline workflows to optimize chemotherapy safety, an approach supported by broad panels such as WES.

We acknowledge several limitations in our study. Although the cohort is sizable, it may not fully capture the genetic diversity of all Indian populations. Furthermore, the inherent constraints of next-generation sequencing (NGS), including potential gaps in coverage and challenges in detecting certain variant types (e.g., large structural variants or deep intronic mutations), may limit the identification of all clinically relevant germline variants. Despite these limitations, our findings emphasize the broader clinical utility of germline testing in detecting risks beyond cancer, with direct implications for surveillance, preventive care, and safer chemotherapy through integration of pharmacogenomic insights. They also provide valuable insights into the genetic architecture of hereditary breast cancer in India, underscoring the importance of ancestry-aware testing and interpretation frameworks to guide risk stratification, therapeutic decision-making, and family-based screening.

Conclusions

Our study underscores the value of comprehensive germline testing in Indian breast cancer patients for cancer risk, secondary findings, and pharmacogenomic markers. Given the high rate of actionable findings, even in unselected cohorts, testing should move beyond restrictive criteria toward inclusive, ancestry-aware strategies, enabling early detection, targeted therapy, and familial risk reduction.

Authors’ contributions

JP and JJ coordinated tissue sample collection and performed exome sequencing. AC assisted exome sequencing. SGR and SR coordinated tissue sample and patient follow up information. BJ performed the data analysis, contributed to the study design, and participated in writing the manuscript. RR assisted with data analysis. SS, VS, and SM conceived, designed, coordinated the study, and contributed to manuscript preparation.

Funding

This work was supported by grants from the Institute of Eminence-Centre of Excellence “Centre for Cancer Genomics and Molecular Therapeutics”, Ministry of Education [SP2223/1242/CPETWOCGMHOC] Govt. of India and Indian Institute of Technology Madras, Chennai, India to SM.

Data availability

The datasets generated and/or analyzed during the current study are available in the repository “Bharat Cancer Genome Atlas (BCGA)” [https://www.bcga.iitm.ac.in] (about: blank).In this study, we analyzed a comprehensive collection of 479 frozen tissue samples along with corresponding sample data, sourced from the NCTB at IITM. Sample collection process was rigorously conducted after the Indian Institute of Technology Madras, Chennai Institutional Ethical Committee approvals (IITM/IEC/2014043; IEC/2019-03/SM/01/01). All patients gave written informed consent regarding the collection and storage of cancer tissue samples and the use of these samples for research purposes and publications. The study was conducted in accordance with the 1964 Helsinki Declaration and later amendments.

Declarations

Competing interests

The authors declare no competing interests.