Navigating Genetic Testing in Nephrology: Options and Decision-Making Strategies

Abstract

Technological advances such as next-generation sequencing (NGS) have enabled high-throughput assessment of the human genome, supporting the usage of genetic testing as a first-line tool across clinical medicine. Although individually rare, genetic causes account for end-stage renal disease in 10% to 15% of adults and 70% of children, and in many of these individuals, genetic testing can identify a specific etiology and meaningfully impact management. However, with numerous options for genetic testing available, nephrologists may feel uncomfortable integrating genetics into their clinical practice. Here, we aim to demystify the process of genetic test selection and highlight the opportunities for interdisciplinary collaboration between nephrologists and genetics professionals, thereby supporting precision medicine for patients with kidney disease. We first detail the various clinical genetic testing modalities, highlighting their technical advantages and limitations, and then discuss indications for their usage. Next, we provide a generalized workflow for genetic test selection among individuals with kidney disease and illustrate how this workflow can be applied to genetic test selection across diverse clinical contexts. We then discuss key areas related to the usage of genetic testing in clinical nephrology that merit further research and approaches to investigate them.

Although individually rare, genetic etiologies of kidney disease collectively account for end-stage kidney disease in 10% to 15% of adults and 70% of children.^1,2 Genetic and acquired forms of kidney disease can differ substantially in their prognosis, course, and management, but are often challenging to distinguish from each other using traditional diagnostics alone.^3,4 Moreover, because early-stage kidney disease is frequently clinically silent, individuals may present only when they reach end-stage kidney disease, at which point traditional diagnostics such as renal imaging and renal biopsy may be uninformative or disfavored.⁵ Genetic testing can also enable early detection of kidney disease in family members, supporting targeted management and improved outcomes.⁶

Traditionally, geneticists meet with individuals suspected to have a genetic condition and conducted a detailed clinical evaluation, including personal and family history, physical examination, and relevant biochemical, imaging, and/or histopathological testing. Based on the results of this evaluation, focused genetic testing involving 1 or several candidate gene(s) is ordered to confirm the clinically suspected disease. In nephrology, clinicians learn to recognize individuals with features of some monogenic forms of kidney disease, such as Alport syndrome,⁷ and refer them for genetic evaluation. Studies now support that genetic testing has the potential to help achieve early and specific diagnosis and personalized management for many more individuals with kidney disease, including those with unknown etiology of disease.^4,8, 9, 10 In addition to resolving the diagnostic odyssey, a process that carries significant time, financial, and psychosocial costs for patients with rare diseases and their families,^11,12 genetic diagnosis can also prevent individuals from having to undergo further, potentially invasive tests. Moreover, knowledge of the specific etiology of chronic kidney disease (CKD) can inform clinical management,^4,9,13 and help evaluate living kidney donors.¹⁴

Massively parallel sequencing, also known as NGS has revolutionized medical genetics by enabling high-throughput assessment of variations across the human genome.^15,16 Such technological advancements are shifting practice paradigms in clinical genetics.^17,18 NGS-based tests, such as targeted panels or exome sequencing (ES), are deployed as part of the initial diagnostic workup of individuals with suspected genetic disease,^19,20 with clinical evaluation following up on the genetic test results.²¹ Review of NGS-based studies shows that a monogenic cause of a disease can be found in approximately 5% to 30% of adults and 30% of children with CKD, albeit with variable diagnostic yield: higher yields are seen among individuals with a positive family history of CKD, earlier age of CKD onset, and presence of extrarenal features.^4,9,13

In this shifting landscape, nephrologists are faced with a bevy of different options for genetic testing and may feel insufficiently prepared to choose between them.²² In this review, we focus on genetic test selection for rare Mendelian forms of kidney disease and refer those interested in other uses of genetic testing to other relevant reviews.23, 24, 25 We also refer readers interested in approaches for genetic testing for unaffected individuals (e.g., unrelated candidate living kidney donors) to relevant reviews on this topic.^14,26 We first provide a detailed overview of the different genetic testing modalities available, focusing on their analytical advantages, drawbacks and indications for their usage. We then build on previous work on applying genomics for kidney disease^6,27 to outline frameworks for genetic test selection among individuals with kidney disease and discuss how they can be applied across different clinical contexts. Finally, we present key areas for further study in clinical genetic testing in nephrology.

Types of Genetic Variants

Genetic variants are alterations in the DNA sequence relative to the standard (reference) sequence (Figure 1). They can be categorized based on their size (e.g., single-nucleotide variant [SNV] vs. structural variants [SVs]), location (e.g., intronic vs. exonic), impact on the encoded protein (e.g., synonymous vs. nonsynonymous), or clinical significance (e.g., pathogenic vs. benign). These 4 categorizations can be simultaneously applied to any given genetic variant (e.g., a variant can be a single nucleotide change located in an exon, encoding for a synonymous change, and classified as benign). Although the term “mutation” has been used to describe DNA sequence alterations, it has been advised to avoid using this terminology because the term “mutation” implies that a change is disease-causing and has negative connotations (i.e., being popularly associated with the grotesque and disturbing).²⁸

Figure 1.

Genetic technologies for variant detection. Different clinical genetic testing modalities are illustrated with the type/s of genetic variants they can detect. ES, exome sequencing; FISH, fluorescent in situ hybridization; GS, genome sequencing; MLPA, multiplex ligation probe–dependent amplification; NGS, next-generation sequencing.

Here, we group genetic variants into 4 broad categories based on their size spectrum: SNVs, which alter a single DNA base; insertions and deletions (indels), which can be small (< 50 base pairs) or large (≥ 50 base pairs); variable tandem number repeats, which consist of a variable number of DNA sequence repeats in tandem; and SVs, which are alterations of ≥ 1 kilobases of DNA. SNVs encompass single-nucleotide polymorphisms, which are present at frequencies in ≥ 1% of individuals in a population, and rarer single DNA base alterations (present in < 1% of individuals in a population).²⁸ In addition, we discuss epigenetic variants, which are modifications to the DNA molecule or associate proteins that do not alter the DNA sequence but can impact gene expression.

Genetic Testing Methodologies: Strengths and Limitations

Genetic technologies vary in their ability to detect different categories of variants (Figure 1, Table 1, and Table 2). The methodologies have evolved, with older methods often having high technical accuracy but being increasingly time- and cost-inefficient as more genes are associated with diseases.

Table 1.

Characteristics of different clinical genetic testing modalities

Test	Indications for use	Advantages	Limitations	Turn-around time	Price	Examples of usage	Examples of incidental or secondary findings
Sanger sequencing	Validation of NGS findings Regions poorly assessed by NGS, such as GC-rich and/or highly repetitive regions Individuals with features specific to a form of kidney disease with a single gene etiology Screening for previously identified disease-causal variant in familial cascade testing Identifying known recurrent disease-associated variants	High technical specificity and sensitivity Relatively simple sequence interpretation	Inefficient and costly if used to test long genes or multiple genes	2–5 d	$ (assuming single variant testing)	Confirm frameshift variant detected by NGS Detect a glycine substitution variant in the first exon of COL4A3, which has high GC content, in an individual with familial early-onset hematuria and sensorineural hearing loss for whom NGS testing identified no candidate COL4A3/4/5 variants Identifying WT1 variant in an individual presenting with steroid-resistant nephrotic syndrome with diffuse mesangial sclerosis on renal biopsy, Wilm's tumor, and ambiguous genitalia APOL1 G1/G2 risk allele testing	None, as the test solely involves a particular gene/genetic region related to the patient's phenotype
Multiplex ligation probe–dependent amplification (MLPA)	Identifying intragenic CNVs for a single gene or several genes Validation of intragenic CNVs identified by NGS-based CNV calling	High technical accuracy Easy and fast interpretation Can be modified to include methylation testing (MS-MLPA)	Not commercially available for all genes Cannot identify CNVs outside of the genetic regions targeted Cannot detect balanced rearrangements Limited detection of genetic mosaicism	7–14 d	$	Detection of a COL4A5 deletion as part of genetic testing for X-linked Alport syndrome (case S3) Detection of NPHP1 deletion in a child presenting with polydipsia, polyuria, and increased echogenicity with reduced corticomedullary differentiation on ultrasound	If used for validation of NGS-called CNV, CNV within a gene associated with a disorder unrelated to the patient's clinical presentation
Karyotype	Detection of chromosomal aneuploidies and balanced chromosomal rearrangements > 5 Mb Investigation of infertility, recurrent miscarriages, and/or multiple congenital anomalies (because these can be due to balanced chromosomal rearrangements)	Detects both balanced (e.g., balanced translocations, inversions) and unbalanced (e.g., deletions, duplications) structural variants	Cannot detect uniparental disomy Requires mitotically active (dividing) cells and cell culture	≤ 14 d	$–$$	Confirm diagnosis of Turner syndrome in an individual with short stature, primary amenorrhea, and horseshoe kidney Detect balanced Robertsonian translocation involving chromosome 21 in an individual with a history of multiple miscarriages and a child with trisomy 21 (Down syndrome) with posterior urethral valves29,30 Detected de novo balanced translocation in a fetus with bilateral renal agenesis31	Sex chromosome aneuploidy Balanced rearrangements (e.g., inversions or translocations)
Fluorescent in situ hybridization (FISH)	Investigation for structural variants in known disease-associated genomic regions (targeted by the probe), including microdeletions and microduplications, chromosomal aneuploidies, and balanced rearrangements Identify mosaic structural variants Follow up after chromosomal microarray to give positional information, which can help determine recurrence risk	Detects both balanced (e.g., balanced translocations, inversions) and unbalanced (e.g., deletions, duplications) structural variants in the genomic regions targeted Can detect mosaic structural variants	Cannot detect uniparental disomy Can only detect variants in the genomic regions targeted	≤14 d	$$–$$$ (depending on number of regions targeted)	Prenatal screening for common aneuploidies, such as those involving gains or losses of chromosomes 13, 18, 21, X, and Y Identify 17q12 deletion in a patient with cystic dysplastic kidneys, hypomagnesemia, and intellectual disability Detection of mosaic 22q11.2 deletion (DiGeorge syndrome) in an individual with cleft palate, failure to thrive, and tetralogy of Fallot32 Identification of a maternally inherited balanced translocation involving chromosomes 19 and 21 on prenatal screening of a family with a history of multiple miscarriages	None
Chromosomal microarray (CMA)	Individuals with CAKUT with or without extrarenal anomalies (case S2) Individuals with multiple congenital anomalies and/or neurodevelopmental delay	Genome-wide CNV detection	Limited technical sensitivity for low-grade genetic mosaicism, and CNVs in regions that are more difficult to capture using array technology (e.g., repetitive elements, pseudogenes) Cannot provide positional information (e.g., whether a duplication is located in tandem [next to] the original copy of the gene or at another site in the genome; position can be determined by karyotyping or FISH)	2–3 wks	$$	Identify 22q11.2 deletion (DiGeorge syndrome) in an individual with unilateral renal agenesis, cleft palate, and immunodeficiency Identify 16p11.2 duplication in an individual with autism, developmental delays, and posterior urethral valves	CNV within a gene associated with a disorder unrelated to the patient's clinical presentation.
Targeted NGS panels	Individuals whose presentation is consistent with a specific clinical subtype of kidney disease Clinical subtypes of kidney disease that have relatively low genetic and/or phenotypic heterogeneity	Can be engineered to provide high coverage of variants in targeted regions, increasing technical sensitivity Assessment of a selected set of genes (i.e., genes associated with a particular clinical subtype of kidney disease) facilitates variant interpretation and reduces risk of secondary genetic findings	Limitations inherent to panel design, including sensitivity-specificity tradeoff for gene selection, inter-laboratory variation in which genes are chosen, and needing to frequently review and update panel (e.g., to add newly identified genes and/or remove genes found to have insufficient evidence of disease association) Low capacity for reanalysis of the genetic data generated Increased potential for VUS to be returned33	10–21 d	$–$$	Testing complement pathway–associated genes in a patient with atypical hemolytic uremic syndrome34 Testing for genes associated with ADPKD in an individual presenting with bilateral enlarged kidneys with multiple cysts and a negative family history (case S1)	None, because the panel is targeted to the patient's specific subtype clinical disease
Broad NGS panels	Clinical presentation associated with a wide range of genetic causes	Can be engineered to provide high coverage of variants in targeted regions, increasing technical sensitivity Assessment of most genes known to be associated with kidney diseases	Limitations inherent to panel design, including sensitivity-specificity tradeoff for gene selection, interlaboratory variation in which genes are chosen, and needing to frequently review and update panel (e.g., to add newly identified genes and/or remove genes found to have insufficient evidence of disease association) Low capacity for reanalysis of the genetic data generated Very high potential for VUS to be returned, than targeted NGS panel or ES/GS.33	10–21 d	$–$$	Patient with a category of kidney disease that has high genetic heterogeneity Patient with kidney disease of unknown etiology (case S4)	For broad NGS panels that target a general clinical disease category (e.g., monogenic kidney diseases), there can be incidental findings in genes associated with diseases that are associated with the general clinical disease category but are unrelated to the patient's clinical disease subtype
Exome sequencing (short-read)a (ES)	Individuals with clinical subtypes of kidney disease that have high genetic heterogeneity Individuals with nonspecific clinical features, including kidney disease of unknown etiology Individuals left undiagnosed by more targeted testing	Genome-wide testing provides higher diagnostic sensitivity than targeted approaches (e.g., single-gene or multigene panel testing) Assessment of the coding regions, which harbor the majority of known disease-causal variants, increases the efficiency of genome-wide analysis than genome sequencing Interrogation of all protein-coding regions supports sequence reanalysis in the case of novel gene discovery Option to receive medically actionable secondary genetic findings	Limited coverage of certain regions and lower detection accuracy for certain types of variants (e.g., indels) than genome sequencing Misses intronic regions unless specifically targeted by the exome capture kit Genome-wide testing can yield multiple candidate variants, increasing the time needed for analysis and follow-up evaluations	2–3 mo	$$	Genetic diagnostic evaluation of individuals with steroid-resistant nephrotic syndrome.35 Identifying a pathogenic TMEM67 variant [conveying the genetic diagnosis of NPHP-RC] in an individual with presumed ARPKD.36 Identifying a cause of disease among individuals with kidney disease of "unknown etiology" (Table 4).35	Medically actionable secondary findings, e.g., pathogenic or likely pathogenic variants in the ACMG medically actionable genes.37
Genome sequencing (short-read)a (GS)	Individuals with clinical subtypes of kidney disease that have high genetic heterogeneity Individuals with nonspecific clinical features, including kidney disease of unknown etiology Individuals with nondiagnostic results on previous, more targeted testing	Increased technical and diagnostic sensitivity than ES, reflecting its superior per-base coverage and ability to assess variants in both the coding and noncoding regions Higher sensitivity for structural variants than ES Option to receive medically actionable secondary genetic findings	Generates a large amount of data, conferring substantial interpretative and financial costs (e.g., time for interpretation, cost of data storage, etc.) Noncoding variants can be often challenging to interpret and require further follow up (e.g., functional assays, clinical phenotyping) to help clarify their significance	2–3 mo	$$$	Identifying disease-causal intronic variants, e.g., in an intronic DGKE variant in case of genetically unresolved atypical hemolytic uremic syndrome.31 Genetic testing for ADPKD (because of high sequence homology of PKD1).38 Detection of pathogenic balanced structural rearrangements in cases of multiple congenital anomalies.31	Medically actionable secondary findings, e.g., pathogenic or likely pathogenic variants in the ACMG medically actionable genes.37 CNV within a gene associated with a disorder unrelated to the patient's clinical presentation

ACMG, American College of Medical Genetics and Genomics; ADPKD, autosomal dominant polycystic kidney disease; ARPKD, autosomal recessive polycystic kidney disease; CAKUT, congenital anomalies of the kidney and urinary tract; CNV, copy number variant; ES, exome sequencing; GS, genome sequencing; MLPA, Multiplex ligation probe–dependent amplification; MS-MLPA, methylation- Multiplex ligation probe–dependent amplification; NGS, next-generation sequencing; NPHP-RC,nephronophthisis-related ciliopathies; VUS, variant of uncertain significance.

As the absolute cost of genetic testing can vary substantially due to multiple factors, including international differences in healthcare policy, here we only provide the relative costs of the genetic tests discussed, with the number of dollar signs ($) corresponding to higher cost.

^aCurrently, clinical exome sequencing (ES) and genome sequencing (GS) technologies use short-read sequencing, which typically assesses paired DNA segments of 150 base pairs in length. We discuss long-read sequencing approaches in the section “Emerging genetic technologies”.

Table 2.

Detection capacities of different genetic testing methodologies

Genetic testing methodology	Number of genes or genetic regions analyzed	Methylation	Noncoding variants	SNV	Small Indels	Balanced structural variants	Unbalanced structural variants	Possibility of secondary or incidental finding/s	Genetic mosaicism detection
Sanger sequencing	1 gene	No	Yes	Yes	< 10 bp	No	No	None	Limited
MLPA	< 50 genetic regions	No	Yes	No	No	No	1 kb–15 Mb	Low	No
Karyotype	All	No	Yes	No	No	Yes	> 5 Mb	1–3%	No
FISH	1	No	Yes	No	No	Yes	> 100 kb	None	Yes
Chromosomal microarray	genome-wide	No	Yes	No	No	No	> 50 kb	1–3%	No
Methylation array	Range of 1 to genome-wide	Yes	Yes	No	No	No	No	Low	Limited
Targeted NGS panels	5–250	No	No	Yes	< 100 bp	No	Variablea	Low	Yes
Broad NGS panels	350–500	No	No	Yes	< 100 bp	No	Variablea	Variablea	Yes
Exome sequencingb	genome-wide	No	No	Yes	< 100 bp	No	Variablea	∼3%	Yes
Genome sequencingb	genome-wide	No	Yes	Yes	Yes	No	Variablea	∼3%	Yes

FISH, Fluorescent in situ hybridization; MLPA, Multiplex ligation probe–dependent amplification; NGS, next-generation sequencing; SNV, Single nucleotide variants.

^aDepending on laboratories and often require confirmation with FISH, chromosomal microarray or MLPA.

^bShort-read sequencing.

Genetic Testing for SNVs and Small Indels

Sanger Sequencing

For decades, Sanger sequencing was the main methodology to identify SNVs and small indels (< 50 base pairs), which comprise the majority of known causal variants for monogenic forms of kidney diseases (Figure 1). Although Sanger sequencing still has the highest technical sensitivity and specificity for detecting SNVs and indels,³⁹ its limited scope and low throughput have led to a reduction of its utilization in clinical genetic testing. Currently, Sanger sequencing is typically used for the analysis of 1 or several specific variant(s), such as the detection of founder variants (e.g., the APOL1 G1 and G2 risk variants) or of a known variant in the setting of familial testing (Table 1 and Table 2). Analysis of 1 or several specific variants can also be performed using targeted genotyping approaches, such as the Taqman assay.⁴⁰ Because it is a relatively inexpensive and high-throughput approach, the Taqman assay has been applied to detect known kidney disease risk variants (e.g., the APOL1 G1 and G2 risk alleles). It is also a potential testing strategy for screening for founder variants, particularly in resource-limited settings.⁴¹

Short-Read NGS

NGS enables simultaneous sequencing of multiple segments of DNA, facilitating high-throughput assessment of the human genome. NGS has therefore superseded Sanger sequencing in clinical genetics. NGS can detect SNVs with very high sensitivity and specificity, and indels with relatively high sensitivity and specificity. NGS can also identify SVs using bioinformatics tools; however, its sensitivity and specificity are lower than for SNVs and are highly variable.⁴² Although long-read sequencing methodologies are now starting to be used for rare disease diagnosis (Table 3),^43,44 clinical NGS is currently based on short-read sequencing, which usually assesses paired DNA segments of 150 base pairs in length.¹⁶ For clinical evaluations, the predominantly used tests are targeted and broad panels, ES, and genome sequencing (GS).

Table 3.

Emerging genetic technologies

Although the clinical adoption of chromosomal microarray (CMA) and next-generation sequencing (NGS) has revolutionized the diagnosis of rare genetic conditions, approximately 50% to 60% of individuals with suspected Mendelian disease remain unresolved after evaluation with such technologies.17,45, 46, 47 For such individuals left unresolved by standard clinical genetic testing, emerging technologies such as optical genome mapping (OGM), long-read sequencing (LRS), DNA methylation analysis, and RNA sequencing may help achieve a genetic diagnosis, supporting personalized care.

OGM uses imaging of fluorescently labeled mega base–long DNA molecules to identify structural variants (SVs) genome-wide.48,49 OGM can detect copy number variants (CNVs) less than 25kb in size and balanced rearrangements (e.g., inversions or translocations), which are missed by CMA or NGS, as well as somatic mosaicism. Given its ability to detect a variety of SVs, it has the potential to become a first-line approach for prenatal cytogenetic testing.50 OGM has been applied to detect such CNVs across individuals with a variety of rare genetic diseases. It identified a mosaic de novo deletion and inversion of the CDKL5 gene in an infant with epilepsy who had previously undergone CMA, targeted panel, exome sequencing (ES), and genome sequencing (GS) testing.51 OGM also identified a 2.8kb insertion within the SMARCB1 gene in 2 siblings with atypical teratoid rhabdoid tumor left unresolved by ES with CNV analysis.52 Currently, there is a clinically approved OGM-based assay for identifying the 3.3kb D4Z4 repeat array causal for facioscapulohumeral muscular dystrophy (FSHD), which has the potential to replace the traditional time- and labor-intensive Southern blot for FSHD diagnosis.53 Although OGM offers the ability to detect both balanced and unbalanced SVs of varying sizes in a single test, it has important limitations: it requires high molecular-weight DNA, which is more difficult to extract, does not provide information about the sequence of the SV(s) detected, and its resolution is limited by the spacing of the fluorescent labels along the DNA molecule and the resolution of the imaging photocell.49,54

LRS refers to sequencing that yields continuous DNA or RNA reads of between 1kb and 4 megabases.55 LRS can be targeted to a specific genomic region (i.e., T-LRS) or done across the genome (lrGS). Current commercial LRS technologies include PacBio HiFi sequencing, which uses DNA polymerase–mediated incorporation of fluorescently labeled nucleotides into a circular DNA template to identify the DNA bases present in a particular genomic region; and Oxford Nanopore Technologies (ONT) sequencing, which analyzes changes in electrical current as a single-stranded DNA or RNA molecular travels through a protein nanopore.56 Although ONT and PacBio sequencing vary in their read lengths and per-base error rates, both offer substantially longer sequence reads than the reads from standard (short-read) NGS. These longer reads enable more accurate detection of variants that can be missed by short-read NGS, such as variants occurring in highly repetitive regions of the genome and repeat expansion variants. LRS also allows for phasing of variants without the need for parental samples, which can help ascertain the variants’ pathogenicity. In addition, both technologies can detect alterations in DNA methylation, enabling simultaneous evaluation of the sequence and methylation pattern at a particular genomic region.57,58 Recent research-level sequencing studies highlight the potential of T-LRS and lrGS for cases left undiagnosed by short-read NGS. As a targeted approach, T-LRS can be used to identify disease-causal variants among genetically unresolved individuals with specific features for a monogenic form of disease. In nephrology, T-LRS led to the identification of the “missing” second variant in WDR19 in an individual with nephronophthisis, retinal dystrophy, skeletal dysplasia, and hepatic fibrosis in whom an NGS ciliopathy gene panel had identified a single heterozygous pathogenic WDR19 variant, thus genetically confirming the diagnosis of Senior-Loken syndrome.59 T-LRS also allowed the identification of intronic variants in individuals with genetically unresolved Gitelman syndrome.60 T-LRS has also been successfully applied for the genetic diagnosis of autosomal dominant polycystic kidney disease, reflecting its superior detection sensitivity than short-read NGS and multiplex ligation probe–dependent amplification to detect variants in genes such as PKD1 that have high sequence homology.61 As a genome-wide approach, lrGS has been shown to provide a molecular diagnosis among pediatric patients with a variety of rare disease phenotypes who had been left undiagnosed by previous genetic testing, reflecting its superior ability to detect SVs and variants in regions poorly captured by short-read NGS.62 Although LRS has increased technical sensitivity than short-read NGS, as a novel technology, there is a need for standardized analytic pipelines and larger population reference datasets, which can make variant filtering and prioritization challenging. Ongoing initiatives using LRS, such as the All of Us project63 and the 1000 Genomes project64 will help address these challenges. In addition, though recent investigations support the diagnostic utility of LRS to resolve cases left undiagnosed by short-read NGS, systematic large-scale comparisons of the diagnostic yield of LRS versus existing genetic testing modalities are needed to inform clinical usage of LRS.

DNA methylation analysis can be performed using methylation arrays or LRS. Certain genetic diseases, including some syndromes involving neurodevelopmental delay and multiple congenital anomalies, result from altered DNA methylation (DNAm).65 These include syndromic disorders that can present with congenital anomalies of the kidney and urinary tract, such as Kabuki syndrome66 and CHARGE syndrome.67 Because these conditions have unique DNAm patterns, known as “episignatures,”DNAm analysis can help achieve a molecular diagnosis for affected individuals.68 Traditionally, targeted DNAm analysis was performed for individuals with features of a specific imprinting disorder to confirm the clinically suspected diagnosis.69 Now, genome-wide DNAm analysis emerging as a follow-up test to resolve individuals with variants of uncertain significance (VUS) in genes associated with epigenetic disorders: here, identification of an episignature for a specific disorder supports that the individual has the condition, and can help characterize the functional impact of the VUS (case S2).70 However, there are important limitations. Although DNAm analysis is typically performed using peripheral blood samples, there is considerable tissue-specific variation in DNAm patterns, limiting the generalizability of the results observed from blood samples.71 In addition, there is the need to develop consensus technical standards for DNAm testing, because analysis must include correction for factors that can affect DNAm patterns, such as age, sex, and environment.70 Moreover, existing methylation arrays have a limited ability to detect low-grade genetic mosaicism and can only identify disorders that have known episignatures, lowering the diagnostic sensitivity of DNAm testing.68 Nonetheless, DNAm analysis has been recently reported to resolve as many as 30% of individuals with features suggestive of a rare neurodevelopmental disorder,72a diagnostic yield comparable to those reported for CMA (15%–20%) and ES (30%–40%).73 These findings support further investigation of the diagnostic utility of DNAm analysis, particularly for individuals with features suggestive of epigenetic disorders and/or with a VUS in a gene associated with a disease that has an episignature.

RNA sequencing (RNA-seq) applies massively parallel sequencing methods to sequence and quantify the different RNA molecules (i.e., the transcriptome) within a cell. RNAseqcan therefore be used to evaluate whether a DNA sequence variant in a given gene impacts RNA splicing and expression of that gene, which can help determine the variant’s pathogenicity.74 For example, in an analysis of nearly 700,000 individuals who underwent clinical NGS for different genetic diseases, incorporating RNA-seq helped resolve 31% of the VUS identified by clarifying their splicing impact.75 Similarly, RNA-seq of individuals with autosomal dominant polycystic kidney disease with deep intronic VUS detected on GS validated that these variants altered splicing and led to reduced PKD1 gene expression, supporting their pathogenicity.76 In addition to resolving VUS, RNA-seq can identify candidate disease-causal genes by identifying genes with altered expression in affected individuals relative to controls.77 However, there are important caveats. Both individual and experimental factors (e.g., age, sequencing platform) can impact transcriptome profiles, requiring the use of in-house control sets to help ensure accurate analysis. Moreover, because the transcriptome varies substantially between different tissues, it is crucial that samples be taken from the tissue(s) relevant to the disease of interest. Recent studies in mouse models demonstrate that genes associated with monogenic forms of kidney disease showed cell-specific expression patterns, supporting further investigation of the clinical utility of single-cell–based RNA-sequencing nephrology.78

Targeted and Broad NGS Panels

Genetic panels analyze a predetermined set of genes associated with a clinical phenotype, with the number of genes assessed reflecting the scope of phenotypes included (Figure 1, Table 1 and Table 2). Targeted panels assess a specific clinical subtype of kidney disease and generally range from 3 to 20 genes; for example, a panel for Alport syndrome consisting of COL4A3, COL4A4, COL4A5, and COL4A6 genes. Broad panels can cover a wider category of kidney diseases that may include multiple specific clinical subtypes; for example, a panel for steroid-resistant nephrotic syndrome that includes 50 genes associated with hereditary podocytopathies.⁷⁹ Of note, there are also broad panels that include multiple forms of monogenic kidney disease that contain hundreds of genes.^80,81 Genetic reports for panels usually include the list of the genes included and the limitations of the test (e.g., exons not sequenced or types of variants not analyzed).

ES

ES refers to the targeted capture of most of the protein-coding regions in the genome (i.e., the exome), allowing a broad search for pathogenic variants in the absence of clinical clues pointing to a specific genetic etiology. ES has highly variable technical sensitivity across the exome, with poor detection capacity for certain clinically relevant regions (Figure 1, Tables 1 and 2).⁸² ES can thus have lower diagnostic sensitivity for some variants, for example, in PKD1 regions with high sequence homology.⁸³ In addition, ES capture kits can vary in their sequencing coverage for different genes. For example, coverage for the gene GREB1L, associated with renal agenesis, varies highly across different ES capture kits.⁸⁴ Generally, ES reports do not provide detailed information on the exon-level sequencing coverage for different genes. Thus, for individuals with phenotypes consistent with a genetic kidney disease who have negative ES results, we recommend obtaining this information to help rule out these being false negative results because of insufficient sequencing coverage.

GS

In addition to assessing the noncoding regions, GS provides uniform coverage across the protein-coding regions^85,86; thus, it is expected to have greater diagnostic sensitivity than ES. This is especially advantageous for the investigation of genes that are more difficult to accurately assess via ES, such as PKD1; and GS has been recommended as an effective strategy for the molecular diagnosis of autosomal dominant polycystic kidney disease.^38,87 In addition, GS can include analysis of the mitochondrial genome, enabling unbiased detection of individuals with kidney disease because of mitochondrial DNA variants,⁸⁸ further increasing its diagnostic sensitivity.⁸⁹

Genetic Testing for SVs

Karyotyping

Karyotyping was the first technology developed to visualize SVs, and remains a first-line method for the detection of large SVs, including chromosomal aneuploidies, translocations, inversions, and complex rearrangements.^90,91 Examples of usage of karyotyping in nephrology include identifying balanced translocations among individuals with syndromic and isolated congenital anomalies of the kidney and urinary tract and detecting complex rearrangements among individuals with Wilms tumor-aniridia-genital anomalies-retardation syndrome (Table 1).⁹²

Fluorescent In Situ Hybridization

In the 1990s, to better identify microdeletions and certain recurrent translocations, fluorescent in situ hybridization (FISH) was developed and remains commonly used in molecular diagnostics. Examples of usage of FISH in nephrology include identifying known disease-causal SVs, such as the 22q11.2 deletion associated with DiGeorge syndrome, the NPHP1 deletion in nephronophthisis,⁹³ and SVs involving the TFE3 and TFEB genes in renal cell carcinoma.⁹⁴

Multiplex Ligation–Dependent Probe Amplification

In the early 2000s, to increase the number of copy number variants (CNVs) analyzed simultaneously and reduce the cost of CNV detection, multiplex ligation-dependent probe amplification (MLPA) was developed and introduced in clinical care.⁹⁵ The sequences targeted by MLPA are 50 to 70 nucleotides long, allowing MLPA to identify single exon CNVs that are too small to be detected by FISH. In addition, MLPA allows the detection of CNVs for up to 50 genetic regions. However, MLPA is not available for all genes and does not provide information on the size of the deletion. Examples of clinical utilization of MLPA in nephrology include detection of pathogenic deletions involving COL4A5 among individuals with X-linked Alport syndrome⁹⁶ and SVs involving the CFH-CFHR genomic region among individuals with atypical hemolytic uremic syndrome.⁹⁷

Chromosomal Microarray

Karyotyping, FISH, and MLPA have been largely replaced by chromosomal microarray (CMA) as the first-tier test for individuals with phenotypes that can result from pathogenic CNVs, such as multiple congenital anomalies or neurodevelopmental disorders.⁹⁸ Here, we use the term CMA to include 2 different technologies: comparative genomic hybridization arrays and single-nucleotide polymorphism arrays, which can be used to identify unbalanced SVs (Figure 1, Table 1, and Table 2).⁹⁹ Unlike FISH or MLPA, which require a priori selection of genomic regions to target, CMA offers genome-wide CNV assessment, increasing diagnostic sensitivity, particularly for individuals who present with features that are not specific to a single genomic disorder. In nephrology, CMA detected pathogenic CNVs in approximately 10% of individuals with isolated and syndromic congenital anomalies of the kidney and urinary tract.¹⁰⁰ Moreover, CMA has been reported to identify disease-causal CNVs in 5% of children with CKD because of causes other than congenital anomalies of the kidney and urinary tract and in 1.1% of adults with all-cause CKD, suggesting that CNVs may contribute more broadly to CKD than traditionally thought.¹⁰¹

NGS-Based SVs Identification

Although CMA and MLPA remain the gold standards for identifying CNVs, recent studies suggest that CNVs can be detected with high sensitivity using bioinformatic tools from targeted panels’ data; however, their overall performance varies, with lower accuracy reported for single-exon CNVs and CNVs in regions that are more poorly captured by short-read NGS.^102,103 CNVs can be called using bioinformatic methods from ES data. However, GS has superior SV detection ability, and it can identify balanced SVs as well as CNVs.⁹⁰ Considering that laboratories vary in whether they perform these analyses, it is important to check the types of variants analyzed when ordering a test.

Practical Considerations for Choosing Between Different Testing Modalities

When choosing between different genetic testing modalities, in addition to diagnostic sensitivity, other key considerations include cost, turn-around time, and incidental and secondary findings (Table 1).

Short-Read NGS Limitations

Short-read NGS has limited capacity to identify variants in certain parts of the genome, such as highly repetitive regions (e.g., the complement factor H gene cluster)¹⁰⁴ and regions with high sequence homology (e.g., the PKD1 gene and its pseudogenes).¹⁰⁵ Short-read sequencing can have limited coverage of certain regions and lower detection sensitivity for certain variants (e.g., indels). To help address these issues, targeted or broad panels are frequently supplemented with other technologies.¹⁰⁶

The detection of certain variants can be difficult across technologies. For example, in MUC1-associated autosomal dominant tubulointerstitial disease, the most commonly known disease-causal variant is an insertion inside a variable tandem number repeat unit of 60 nucleotides (60-mer), which leads to a frameshift and premature truncation of the encoded protein. Considering that Sanger sequencing and standard short-read NGS analyses are unable to detect this variant, identifying it requires long-range polymerase chain reaction and molecular cloning.¹⁰⁷ The variant’s initial identification and alternate approaches, including computational pipelines using short-read NGS^108,109 and long-read sequencing¹¹⁰ have been developed to identify this variant and may be implemented in the future among individuals with suspected MUC1-associated autosomal dominant tubulointerstitial disease. Owing to such technical limitations, laboratories often supplement targeted panels with other methodologies to increase their detection sensitivity.¹⁰⁶ Recently, new bioinformatics tools to identify variable tandem number repeats from ES and GS have been reported, which have variable detection accuracy.^111,112

Variants of Uncertain Significance

A common concern about genetic testing is the possibility of variants of uncertain significance (VUS). Given that most genetic diseases are rare, current guidelines for clinical variant interpretation assess a variant’s frequency in the general population versus the disease’s known prevalence to help determine its pathogenicity.¹¹³ However, because of multiple factors, including disparities in access to infrastructure and funding, non-European populations have been underrepresented in clinical settings and genomic research such that current population genetic databases do not reflect the diversity of the global population.¹¹⁴ This underrepresentation complicates variant interpretation, because it can be difficult to gauge the true population prevalence of variants, and leads to higher rates of VUS among non-European individuals undergoing clinical genetic testing.115, 116, 117 Importantly, there are ongoing initiatives to bring genomics to historically underrepresented regions, such as the Human Heredity and Health in Africa Kidney Disease Research Network,¹¹⁸ Deciphering Diversities - Renal Asian Genetics Network,¹¹⁹ and Brazilian Network of Pediatric Nephrotic Syndrome.¹²⁰ These efforts will help improve our understanding of the genetics of kidney disease across different populations, reducing the current ancestry-based disparities in VUS rates and supporting individuals of diverse backgrounds benefiting equitably from genetic testing.

Per current guidelines, VUS should not inform clinical decision-making.¹¹³ As genome-wide tests, ES and GS may be expected to return more VUS than targeted panels. However, a recent analysis of over 1.5 million genetic tests from 19 clinical laboratories found a higher VUS rate from targeted panels than from ES or GS.³³ This is because of the differences in VUS reporting protocols; currently, panels generally report all VUS, whereas, for ES or GS, laboratories report only highly compelling VUS (e.g., those associated with genetic conditions that have high clinical overlap with the patient’s listed phenotype and/or segregate appropriately with the disease in a family). Thus, the higher VUS rates reflect the lower specificity for a particular clinical subcategory of disease.³³ Given that resources needed to allow potential VUS reclassification are often limited, the VUS included on a genetic test report should be carefully reviewed (e.g., by a genetics professional) to help determine which ones merit further investigation.

Secondary Findings

ES and GS can also prompt concern about secondary findings. Although unrelated to the primary testing indication (i.e., kidney disease), secondary findings can enable presymptomatic diagnosis and prevention of medically actionable conditions,¹²¹ and may have meaningful implications for kidney care as well.¹²² Given that individuals vary in their preference to receive pretest counseling, it is important to discuss the pros and cons of analysis for these secondary findings and note that individuals can choose to opt out of receiving secondary findings. When choosing between different NGS tests, financial status and insurance coverage should be considered, because insurance coverage for ES or GS is highly variable.^123,124

The American College of Medical Genetics and Genomics currently recommends that individuals undergoing genomic sequencing (i.e., ES or GS) be assessed for medically actionable secondary genetic findings, even if unrelated to the primary indication for genetic testing.^125,126 The current set of American College of Medical Genetics and Genomics medically actionable genes includes multiple genes associated with adult-onset conditions, such as hereditary cancer predisposition syndromes and cardiomyopathies.³⁷ Conversely, the European Society of Human Genetics recommends a cautious approach to such “opportunistic” screening, especially among minors.¹²⁷ In addition to the ethically complex question of whether to disclose such secondary findings for adult-onset conditions to a pediatric patient, analysis of trio (parents and child) ES or GS data for secondary genetic findings may also identify a parentally inherited variant, leading to parents receiving unprecedented information about their own risk for cancer, cardiovascular disease, or other disorders.¹²⁸ Current guidelines by the American College of Medical Genetics and Genomics, the European Society of Human Genetics, and the American Academy of Pediatrics recommend pretest counseling regarding the pros and cons of receiving such secondary genetic findings and providing parents with the option to “opt out” of the investigation and reporting of secondary findings.^129,130

Laboratory Practices

Considering that each laboratory chooses which genes to include in their targeted or broad panels, panels vary in their gene content and thus, their diagnostic sensitivity and specificity for the associated phenotype. Developing a consensus list of genes associated with monogenic forms of kidney disease can help address this interlaboratory variability^6,27; and global initiatives, such as the Gene Curation Coalition,¹³¹ are contributing to this goal by assessing the clinical validity of gene-disease associations. In addition, to increasing the test’s sensitivity, some laboratories supplement their tests with additional testing methodologies.

Although the advent of consensus guidelines for clinical variant interpretation led to more consistent interpretation across different clinical laboratories, there remains notable interlaboratory variations, affecting the tests’ sensitivity.^113,132,133 When choosing a test, it is thus critical to consider the laboratory’s expertise in classifying variants in the genes of interest. Laboratories also vary in follow-up testing provided. For example, some laboratories offer testing of family members to help with the potential reclassification of VUS, which is especially relevant if the patient has affected family members. Some laboratories also offer free testing of family members in case of positive genetic findings (i.e., cascade screening). In addition, some laboratories offer virtual panels where analysis of a selected set of genes is performed on ES or GS data, enabling a broader analysis without requiring additional patient DNA samples if the virtual panel testing yields negative results.

Cost

The price of a genetic test reflects not only the costs of sequencing but also those related to processing, storing, and analyzing the resulting genetic data. Broader tests generate larger amounts of data, which often results in a need for more sophisticated bioinformatic tools and greater data storage capacities. This larger amount of genetic data also increases the time and labor needed for variant interpretation. GS generates the largest amount of data that can confer substantial financial and interpretative costs, including for noncoding variants, which can be difficult to interpret and require further clinical and/or experimental follow-up to help ascertain their significance.¹³⁴ Nevertheless, investigations of children with suspected genetic diseases suggest that GS might be a more cost-effective tool than ES because of its higher diagnostic sensitivity.^135,136 Accordingly, GS is anticipated to become a first-line tool for clinical genetic testing in nephrology in the future.¹³⁷ Practically, the out-of-pocket cost (i.e., the amount paid by the patient) of genetic testing depends on several factors, including the test modality and insurance coverage.¹³⁸ Some laboratories promise a maximum out-of-pocket price for their tests to alleviate the fear of “surprise bills.” Importantly, coverage for genetic testing varies substantially between countries, reflecting international differences in healthcare policy.¹³⁹ Access and costs depend on the test and country requirements, like preauthorization or pretest genetic counseling.¹⁴⁰ Given this variability, we only provide the relative costs of the genetic tests discussed here to each other (Table 1). In the United States, insurance coverage is usually lower for genome-wide testing than more targeted approaches.^124,141,142 For example, insurance coverage for CMA and ES remains variable, with a recent study of 801 pediatric patients with suspected genetic disease reporting that CMA and ES were covered by insurance in 89.6% and 83.4% of patients, respectively.¹²⁴ Although single-payer healthcare systems can help address such uneven coverage of genetic testing, they require deciding how to equitably allocate genetic testing in the setting of finite budgets.¹⁴³ To support this goal, frameworks for the prioritization of genetic testing have been proposed in Europe.^144,145 Additional studies are needed to assess the most cost-effective approach for genetic testing of individuals with kidney diseases.

Turn-Around Time

In general, turn-around time is faster for more targeted forms of genetic testing such as Sanger sequencing and targeted panels than ES or GS (Table 1 and Table 2), reflecting that because they analyze a smaller proportion of the genome, fewer candidate variants need to be reviewed by the geneticists. Similarly, because these more targeted genetic modalities assess gene(s) related to the primary indication for testing (i.e., the suspected genetic subtype of kidney disease), there is a lower possibility of incidental and secondary findings.

Genetic Test Selection for Individuals with Kidney Disease

In this section, we present a general framework for genetic test selection among individuals with kidney disease (Figure 2). In the featured case study (Table 4, Figure 3), we illustrate how this workflow can be applied to test selection in clinical practice. For readers interested in seeing additional examples across diverse clinical contexts, we provide 4 additional case vignettes in the Supplementary Material (cases S1–S4).

Figure 2.

Framework for genetic test selection among individuals with kidney disease. The workflow illustrated here is described in the main text in the section: “Genetic test selection among individuals with kidney disease.” ES, exome sequencing; GS, genome sequencing; MLPA, multiplex ligation probe–dependent amplification; Neg, negative; NGS, next-generation sequencing; Pos, positive; VUS, variant of uncertain significance.

Table 4.

Case study

Case Presentation

GF is a 42-yr-old male with a history of diabetes and hypertension who presents with end-stage kidney disease (ESKD). He reports that his grandmother, mother, and maternal aunt, all of whom are deceased, reached ESKD before the age of 50 yrs, which was presumed to be secondary to diabetes and hypertension. Serum electrolytes and complement studies are unremarkable, and urinalysis shows bland sediment and low-grade proteinuria. Renal imaging reveals shrunken, hyperechogenic kidneys; given these findings and his advanced kidney disease, renal biopsy is not performed. His 51-yr-old sister, KF, wants to donate her kidney to him. KF reports being in good health. She is normotensive and has no anomalies on renal imaging. KF’s laboratory values show serum creatinine 0.8 (estimated glomerular filtration rate of 84 ml/min per 1.73 m2) and fasting glucose of 75 mg/dl; urinalysis is within normal limits. KF notes their family identifies as African American, and she has read that individuals of Western African ancestry may be at higher risk of developing kidney disease. GF and KF ask if any additional tests could help them understand why GF and other members of their family have kidney disease and whether KF can safely donate a kidney to GF.

Genetic Differential Diagnosis

Although diabetic and hypertensive nephropathy are on the differential, GF’s personal and family history of ESKD before the age of 50 yrs supports considering monogenic etiologies for his kidney disease. Although his clinical presentation is not specific for a particular category of genetic kidney disease, the absence of certain features can help narrow the differential, for example, the negative complement studies disfavoring a typical hemolytic uremic syndrome or other complement-mediated renal diseases. Given the cooccurrence of diabetes and renal disease in his family, HNF1B nephropathy is a possibility: though classically associated with cystic renal disease and diabetes, there is variable expressivity, with some individuals displaying only isolated renal hypoplasia or no structural renal anomalies on imaging.146,147 In addition, given that GF is an African American of Western African ancestry, APOL1 nephropathy should also be considered, because approximately 13% of African Americans harbor 2 APOL1 risk alleles.148 Although the APOL1 risk genotype substantially increases kidney disease risk, a minority of individuals with APOL1 risk genotypes develop kidney disease, suggesting that a “second hit,” such as other genetic variant(s) and/or systemic or environmental factors, may be needed to produce disease. A study of over 15,000 individuals with kidney disease found that nearly 10% of individuals with APOL1 risk genotypes harbored a diagnostic variant for a monogenic form of kidney disease, supporting genetic testing in individuals with suggestive features for monogenic forms of kidney disease even if they are known to have an APOL1 risk genotype.149

Clinical Utility of Genetic Testing

For GF, genetic testing has the potential to pinpoint a specific cause of disease and thereby inform living-related donor selection and posttransplant prognosis and management.26 Next-generation sequencing –based genetic testing has been reported to identify a monogenic cause in approximately 10% of adults with ESKD of unknown etiology.150, 151, 152 Knowledge of the transplant recipient's primary kidney disease can help determine recurrence risk in the allograft and also identify individuals at increased risk of other common posttransplant complications (e.g., malignancy for WT1-associated focal segmental glomerulosclerosis, diabetes among individuals with HNF1B nephropathy), helping guide surveillance and choice of posttransplant therapy. In addition, by ascertaining whether family members harbor the disease-causal variant, genetic testing can help determine eligible living-related donors. If genetic testing includes investigation for the APOL1 risk genotypes and/or other medically actionable disorders,37 these findings may have clinical utility. For example, knowledge of the APOL1 risk genotype status can inform about the recurrence risk in the allograft153 and the donor’s risk of developing CKD postdonation.14

Although not causal for an individual’s renal disease, secondary findings in genes deemed medically actionable can inform management, for example, detection of a hereditary cancer predisposition syndrome informing immunosuppression regimen or identification of a pathogenic variant in KCNQ1, associated with long QT syndrome, prompting avoidance of QT-prolonging medications such as macrolide and quinolone antibiotics and azole antifungals.122 Therefore, individuals should be counseled about the possible utility of receiving medically actionable secondary findings should they choose to pursue genome-wide testing.

Options for Genetic Testing

ESKD of unclear etiology and no specific familial form of kidney disease support a broad approach to genetic testing (Figure 2). GF’s options include a broad targeted panel of genes associated with monogenic forms of CKD, virtual panel testing, or genome-wide testing (i.e., exome sequencing [ES] or genome sequencing). Becauseeach of these tests is a reasonable first-line approach,35 GF and his nephrologist engage in shared decision-making. His nephrologist writes a letter of medical necessity describing how GF has kidney disease of unknown etiology with a family history of kidney disease and the clinical utility of genetic testing in his case and obtains insurance coverage for ES for him.

Results and Next Steps

ES reveals that GF has a pathogenic variant in the HNF1B gene. GF is given a genetic diagnosis of HNF1B nephropathy, which has meaningful implications for his clinical management. Identification of HNF1B nephropathy warrants targeted evaluation and surveillance for extrarenal manifestations, which include liver dysfunction, neuropsychiatric disease, and exocrine pancreatic insufficiency.147,154 Moreover, because affected individuals have an increased risk of diabetes posttransplant, avoidance of tacrolimus and reduced corticosteroid usage in immunosuppression regimens is advised, and, for those with ESKD and diabetes, combined kidney-pancreas transplant can be considered.

In addition to the pathogenic HNF1B variant, GF’s ES report notes that he harbors the APOL1 G1/G1 risk genotype and has a heterozygous pathogenic variant in the LDLR gene, associated with autosomal dominant familial hypercholesterolemia. His sister, KF, is found to be negative for the HNF1B variant, but does harbor the APOL1 G1/G1 risk genotype and LDLR pathogenic variant. She is counseled that, per population-based studies, individuals with APOL1 risk genotypes have an ∼2-fold increased risk of ESKD post donation, with the caveat that her risk may be lower because the lifetime risk of ESKD decreases as donors age.155 Despite harboring the APOL1 risk genotype, she still meets institutional cutoffs for acceptable lifetime risk and decides to proceed with donating her kidney.

As individuals with heterozygous familial hypercholesterolemia, GF and KF are referred to a cardiologist specialized in hereditary dyslipidemias for recommended management, including stringent lipid profile monitoring with a low threshold for initiating pharmacotherapy (e.g., with statins, ezetimibe, or PCSK9 inhibitors).156 In addition, their familial hypercholesterolemia also informs immunosuppression regimen via supporting minimizing corticosteroid use, given corticosteroids’ negative impact on lipid profile.

Figure 3.

Clinical genetic testing and outcomes for case study. The workflow for clinical genetic testing and associated outcomes for the case described in Table 4 (“Case Study”) is shown here. APOL1 G1/G1, Apolipoprotein L1 gene G1 risk variant homozygote; HNF1B, hepatocyte nuclear factor 1 beta gene; LDLR, low density lipoprotein receptor gene.

Informed and Shared Choice

Clinicians and patients should engage in shared decision-making throughout the process of genetic test selection, particularly when the evidence for choosing one type of test versus another is equivocal. In addition, we recommend that clinicians consult with a genetics professional (e.g., a genetic counselor and/or clinical geneticist) when faced with complex situations; for example, deciding on next steps for a patient with features consistent with a genetic cause of kidney disease who has nondiagnostic results on initial genetic testing.⁶

Patient With a Known Familial Genetic Diagnosis

Genetic test selection starts by ascertaining whether the individuals’ biological relatives have a known genetic diagnosis (i.e., a pathogenic or likely pathogenic variant in a gene associated with the relevant clinical subtype of kidney disease, Figure 2). If so, we recommend that clinicians pursue targeted testing based on the type of variant that has been identified (Table 2); Sanger sequencing would be best for SNVs and small indels, whereas for SVs, MLPA, FISH, or CMA could be used depending on the size of the SV and the availability of MLPA or FISH tests for the familial SV.

Patient With a Disease Associated With One Gene

If the individual does not have a known familial genetic diagnosis, the options depend on the patient’s clinical presentation. If the patient’s clinical presentation is consistent with a clinical category of genetic kidney disease associated with a single genetic cause (e.g., Fabry disease or cystinosis), then Sanger sequencing, MLPA, or FISH are generally most appropriate (Tables 1 and 2).

Patient With a Disease Associated With Few or Multiple Genes

If the patient’s condition may have multiple genetic causes, it can be further divided into those associated with several genes (e.g., Alport syndrome, autosomal dominant polycystic kidney disease) or many genes (e.g., steroid-resistant nephrotic syndrome, renal ciliopathies). For conditions that result from several genes, we suggest a targeted panel with CNV analysis, which offers a specific and sensitive way to detect variants in these genes. We recommend the test chosen include CNV analysis, which can be done on NGS data using bioinformatic methods or through supplementation with MLPA or CMA.

Patient With a Nonspecific Disease

If an individual does not have features consistent with a specific clinical category of genetic kidney disease (e.g., end-stage kidney disease of unknown etiology and no known familial genetic diagnosis), a broad targeted panel, ES, or GS are all reasonable first-line approaches, with clinicians and patients engaging in shared decision-making to determine which test best fits the patient’s unique needs (Table 1, Table 2, Table 3).

Genetic Results Interpretation

Results of genetic testing can be diagnostic (positive) or nondiagnostic (negative or inconclusive results).

Diagnostic Results

If a pathogenic or likely pathogenic variant for a genetic disease that explains the patient’s clinical presentation is identified, it is a diagnostic result. Diagnostic genetic findings can inform clinical prognosis and management, including evaluation and surveillance for the disease’s different manifestations and initiation of targeted therapy. In addition, family members should be offered genetic counseling and cascade testing to determine their disease risk status, inform their subsequent clinical management, and their eligibility to be related living donors for future renal transplantation.⁶

Nondiagnostic Results

Nondiagnostic results include negative results and inconclusive results. Negative results are when no variants for genetic diseases relevant to the individual's clinical presentation are reported. Inconclusive results are when there are VUS (i.e., variants that are not currently clearly disease-causal or benign) for genetic diseases that could be relevant to the individual’s clinical presentation.

Negative Results

For a patient initially suspected to have a specific clinical category of genetic kidney disease, clinicians can reassess the patient to determine if other clinical categories of genetic kidney disease are possible; if so, further genetic testing for these other possible kidney diseases would be indicated. If the clinical impression remains consistent with the originally suspected category of genetic kidney disease, we suggest additional targeted genetic testing to identify variants that may have been missed because of the technical limitations of the initial genetic test used. For patients with nonspecific phenotypes, we suggest routine reanalysis of their genetic data, which may identify disease-causal variants missed in the initial analysis, such as variants identified by newer bioinformatic pipelines or variants in newly discovered disease-associated genes.¹⁵⁷ In addition, clinicians can consider referring such patients to research studies for further evaluation.^45,46

Inconclusive Results and Follow-Up for Candidate Variants

For patients with inconclusive results, we suggest consulting with a genetics professional to help inform the next steps regarding the VUS identified.^6,27 If the VUS is deemed compelling, follow-up testing can help resolve its clinical significance. For example, if the VUS is in a gene associated with a genetic disease that has pathognomonic findings (e.g., on physical examination or biochemical testing), identification of these features can provide evidence for the variant’s pathogenicity. Familial genetic testing can also help resolve VUS; for example, by identifying de novo variants or determining whether the variant segregates appropriately with the disease (i.e., is present in affected family members and absent in unaffected family members). In this case, because family members are being assessed for a single, known variant, we recommend using targeted approaches based on the specific type of variant, for example, Sanger sequencing for a SNV or MLPA/FISH/CMA for a deletion. Other molecular analyses can help inform variant classification, such as RNA analysis to determine the splicing impact of deep intronic variants⁶⁰ and DNA methylation testing to assess for a disease’s episignature⁷² (Table 3). In addition, with time, additional data may emerge that helps resolve a VUS, including observations of the variant in other affected individuals and/or in population databases and experimental assays demonstrating whether the variant has a functional impact. Data-sharing initiatives such as GenomeConnect, where patients can securely share their genetic and clinical data, can support patients and clinicians in learning about such additional data and lead to the resolution of VUS.^158,159

The Need for Multidisciplinary Care

Nephrologists are not expected to become expert geneticists. However, to ensure equitable access to genetic testing, all nephrologists should be able to identify patients who may benefit from genetic testing and provide the appropriate referrals. As illustrated by recent reports on the association between genetic education and referral to genetic testing, including genetics education across different stages of medical training will help support this goal.^22,160,161 In addition, there is a growing community of nephrogenetics specialists available to assist clinicians; these include genetic counselors, nephrologists, and geneticists with expertise in the genetics of kidney diseases.^10,140,162, 163, 164, 165 In the United States, the National Society of Genetic Counselors has developed a website to help locate genetic counselors with nephrology expertise (findageneticcounselor.org). Similarly, the European Rare Kidney Disease Center Network has created a resource of nephrogenetics reference centers.¹⁶⁶ Collaboration involving nephrologists and genetics professionals can increase the diagnostic yield of genetic testing,^164,165 supporting its implementation in multiple clinical sites.

Future Directions

As genetic technologies continue to advance, there is a need to create evidence-based guidelines for implementing equitable genetic testing in nephrology.^6,27,167 Although expert panels have issued recommendations for genetic testing for certain disorders, such as Fabry disease¹⁶⁸ and steroid-resistant nephrotic syndrome,¹⁶⁹ comparative studies on tests’ sensitivity, specificity, turn-around time, cost, and impact of secondary findings are needed to create data-driven frameworks for genetic test selection across different clinical subtypes of kidney disease. There is a need to identify the best implementation strategies to ensure equitable access to genetic testing in nephrology, particularly in resource-limited settings.¹⁷⁰ In addition, research is needed to guide the deployment of other types of genetic types, such as pharmacogenetics,¹⁷¹ polygenic risk scores,^25,172, 173, 174, 175 and relevant comorbidities and medically secondary findings.¹²² As the scope and usage of genetic testing in nephrology broaden, tests that provide multiple types of information may become first-line modalities. For example, because GS can simultaneously provide information about SVs, SNVs, and indels in both the coding and noncoding regions and can provide pharmacogenetic information, polygenic risk scores, and information on other comorbidities, it may supersede targeted panels, ES, and CMA.

Because novel gene-disease associations are regularly reported, there is a need to assess the extent to which a given gene is associated with a particular form of kidney disease. Global initiatives such as the Gene Curation Coalition, which includes organizations such as the Clinical Genome Resource Kidney Domain Working Group, Genomics England Panel App, and PanelApp Australia are collaborating to lead efforts to identify genes reported causal for different forms of kidney disease and systematically review and classify their clinical validity.^131,176 For example, the Clinical Genome Resource Kidney Domain Working Group uses a standardized semiquantitative framework to assess the clinical validity of genes associated with a variety of kidney disease subtypes, including congenital anomalies of the kidney and urinary tract, complement-mediated kidney diseases, cystic disease and renal ciliopathies, glomerulopathies, and tubulopathies. Large-scale sequencing studies involving ethnically and clinically diverse patient cohorts should also help to assess gene-disease associations.

Establishing systematic guidelines for variant interpretation is likewise crucial to achieving genomic medicine in nephrology. In addition to determining the clinical validity of different gene-disease associations, there is a growing understanding that accurate variant interpretation requires tailoring variant classification guidelines to specific genes. For example, expert working groups have published recommendations for variant interpretation for Alport syndrome¹⁷⁷ and cystinuria.¹⁷⁸ Analyses of large datasets of patients with well-characterized monogenic forms of kidney diseases will help create sufficiently sensitive and specific gene-specific guidelines and thereby support increased diagnostic yield and reduced rates of VUS on clinical genetic testing.

As variant classification continues to evolve, clinicians and patients must be appropriately counseled regarding the need to periodically reevaluate genetic results, because variant reclassification can meaningfully inform clinical management.¹⁷⁹ Incorporation of artificial intelligence may help by increasing the efficiency of these efforts; however, further study is needed to assess the accuracy of artificial intelligence– based methods in these domains.^180,181 Of note, recent studies in undiagnosed rare disease cohorts highlight that resequencing can identify causal variants that were initially missed because of technical limitations of the original method of genetic testing.^157,182 As genetic technologies advance and sequencing costs continue to decrease, there is a need for further studies to compare the diagnostic yield and cost efficacy of resequencing patient samples versus reanalyzing previously generated sequence data.¹⁸³

Conclusion

Genetic testing is emerging as a first-line diagnostic tool in nephrology, with the potential to enable diagnosis and targeted management in a variety of clinical contexts (Table 4, Figure 3, cases S1–S4). New modalities, such as optical genome mapping, long-read sequencing, DNA methylation analysis, and RNA sequencing optical genome mapping, are now being piloted for rare disease diagnosis, including kidney disease (Table 3). In this review, we aimed to provide nephrologists with a practical framework for genetic test selection and support them in integrating genetics into their clinical practices. As the usage of genetic testing in nephrology continues to grow, we anticipate that multidisciplinary partnerships in clinical practice, scientific research, and medical education will support the development and implementation of evidence-based guidelines that support genetic testing and personalized care for individuals with kidney disease.

Disclosure

All the authors declared no conflicting interests.

Acknowledgments

Funding

EG was funded by the Children’s National Hospital-National Human Genome Research Institute Pediatrics and Medical Genetics Combined Residency Program. HMR is funded by a K01 award, which was awarded by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), one of the branches of the United States National Institutes of Health.

Supplementary Material

Supplementary File (PDF)

Case Vignettes (includes Cases S1–S4). Glossary. Supplementary References.