Extensive experience in statistical modelling, inference, and multiple hypothesis testing by analyzing various types of data, including genomic and gene expression data.
Current Role at Stanford
Primarily collaborating with VA researchers on GWAS using MVP data. Also providing statistical support to some Stanford clinicians conducting investigational research.
Education & Certifications
MS, Stanford University, Statistics (2014)
MS & PhD, Cornell University, Applied Mathematics (1993)
BS, New Mexico Institute of Mining & Technology, Mathematics (1985)
Plasma proteomic signatures of a direct measure of insulin sensitivity in two population cohorts.
The euglycaemic-hyperinsulinaemic clamp (EIC) is the reference standard for the measurement of whole-body insulin sensitivity but is laborious and expensive to perform. We aimed to assess the incremental value of high-throughput plasma proteomic profiling in developing signatures correlating with the M value derived from the EIC.We measured 828 proteins in the fasting plasma of 966 participants from the Relationship between Insulin Sensitivity and Cardiovascular disease (RISC) study and 745 participants from the Uppsala Longitudinal Study of Adult Men (ULSAM) using a high-throughput proximity extension assay. We used the least absolute shrinkage and selection operator (LASSO) approach using clinical variables and protein measures as features. Models were tested within and across cohorts. Our primary model performance metric was the proportion of the M value variance explained (R2).A standard LASSO model incorporating 53 proteins in addition to routinely available clinical variables increased the M value R2 from 0.237 (95% CI 0.178, 0.303) to 0.456 (0.372, 0.536) in RISC. A similar pattern was observed in ULSAM, in which the M value R2 increased from 0.443 (0.360, 0.530) to 0.632 (0.569, 0.698) with the addition of 61 proteins. Models trained in one cohort and tested in the other also demonstrated significant improvements in R2 despite differences in baseline cohort characteristics and clamp methodology (RISC to ULSAM: 0.491 [0.433, 0.539] for 51 proteins; ULSAM to RISC: 0.369 [0.331, 0.416] for 67 proteins). A randomised LASSO and stability selection algorithm selected only two proteins per cohort (three unique proteins), which improved R2 but to a lesser degree than in standard LASSO models: 0.352 (0.266, 0.439) in RISC and 0.495 (0.404, 0.585) in ULSAM. Reductions in improvements of R2 with randomised LASSO and stability selection were less marked in cross-cohort analyses (RISC to ULSAM R2 0.444 [0.391, 0.497]; ULSAM to RISC R2 0.348 [0.300, 0.396]). Models of proteins alone were as effective as models that included both clinical variables and proteins using either standard or randomised LASSO. The single most consistently selected protein across all analyses and models was IGF-binding protein 2.A plasma proteomic signature identified using a standard LASSO approach improves the cross-sectional estimation of the M value over routine clinical variables. However, a small subset of these proteins identified using a stability selection algorithm affords much of this improvement, especially when considering cross-cohort analyses. Our approach provides opportunities to improve the identification of insulin-resistant individuals at risk of insulin resistance-related adverse health consequences.
View details for DOI 10.1007/s00125-023-05946-z
View details for PubMedID 37329449
View details for PubMedCentralID 5866840
Referral patterns for infantile cataracts in two regions of the United States.
Journal of AAPOS : the official publication of the American Association for Pediatric Ophthalmology and Strabismus
BACKGROUND: Delayed treatment of congenital or infantile cataracts can cause deprivation amblyopia. Prompt diagnosis and surgical intervention is critical for optimal outcomes. This study assessed referral patterns for congenital or infantile cataracts in two regions of the United States.METHODS: The medical records of children 0-1 years of age with congenital or infantile cataracts at Stanford University (2008-2018) and Emory University (2010-2015) were reviewed retrospectively.RESULTS: A total of 111 children were included. Of these, 82 (74%) were initially evaluated by a primary care doctor, of whom 40 (49%) were referred directly to a pediatric cataract surgeon. Of 61 newborns 0-2 months of age, 9 (15%) were initially referred to an eye care provider before 6 weeks of age, but the initial evaluation by a pediatric cataract surgeon was delayed until after 6 weeks of age. Referral patterns were similar between the two institutions (P = 0.06).CONCLUSIONS: Many children with congenital of infantile cataracts are initially referred by a primary care doctor to an eye care provider who does not perform pediatric cataract surgery. Nevertheless, the majority of newborn infants with cataracts were evaluated by a pediatric cataract surgeon before 6 weeks of age.
View details for DOI 10.1016/j.jaapos.2021.09.006
View details for PubMedID 34973446
Implicit Bias and the Association of Redaction of Identifiers With Residency Application Screening Scores.
Importance: Diversity in the ophthalmology profession is important when providing care for an increasingly diverse patient population. However, implicit bias may inadvertently disadvantage underrepresented applicants during resident recruitment and selection.Objective: To evaluate the association of the redaction of applicant identifiers with the review scores on ophthalmology residency applications as an intervention to address implicit bias.Design, Setting, and Participants: In this quality improvement study, 46 faculty members reviewed randomized sets of 462 redacted and unredacted applications from a single academic institution during the 2019-2020 ophthalmology residency application cycle.Interventions: Applications electronically redacted for applicant identifiers, including name, sex or gender, race and ethnicity, and related terms.Main Outcomes and Measures: The main outcome was the distribution of scores on redacted and unredacted applications, stratified by applicant's sex, underrepresentation in medicine (URiM; traditionally comprising American Indian or Alaskan Native, Black, and Hispanic individuals) status, and international medical graduate (IMG) status; the application score beta coefficients for redaction and the applicant and reviewer characteristics were calculated. Applications were scored on a scale of 1 to 9, where 1 was the best score and 9 was the worst score. Scores were evaluated for a significant difference based on redaction among female, URiM, and IMG applicants. Linear regression was used to evaluate the adjusted association of redaction, self-reported applicant characteristics, and reviewer characteristics with scores on ophthalmology residency applications.Results: In this study, 277 applicants (60.0%) were male and 71 (15.4%) had URiM status; 32 faculty reviewers (69.6%) were male and 2 (0.4%) had URiM status. The distribution of scores was similar for redacted vs unredacted applications, with no difference based on sex, URiM status, or IMG status. Applicant's sex, URiM status, and IMG status had no association with scores in multivariable analysis (sex, beta=-0.08; 95% CI, -0.32 to 0.15; P=.26; URiM status, beta=-0.03; (95% CI, -0.36 to 0.30; P=.94; and IMG status, beta=0.39; 95% CI, -0.24 to 1.02; P=.35). In adjusted regression, redaction was not associated with differences in scores (beta=-0.06 points on a 1-9 scale; 95% CI, -0.22 to 0.10 points; P=.48). Factors most associated with better scores were attending a top 20 medical school (beta=-1.06; 95% CI, -1.37 to -0.76; P<.001), holding an additional advanced degree (beta=-0.86; 95% CI, -1.22 to -0.50; P<.001), and having a higher United States Medical Licensing Examination Step 1 score (beta=-0.35 per 10-point increase; 95% CI, -0.45 to -0.26; P<.001).Conclusions and Relevance: This quality improvement study did not detect an association between the redaction of applicant characteristics on ophthalmology residency applications and the application review scores among underrepresented candidates at this institution. Although the study may not have been powered adequately to find a difference, these findings suggest that the association of redaction with application review scores may be preempted by additional approaches to enhance diversity, including pipeline programs, implicit bias training, diversity-centered culture and priorities, and targeted applicant outreach. Programs may adapt this study design to probe their own application screening biases and track over time before-and-after bias-related interventions.
View details for DOI 10.1001/jamaophthalmol.2021.4323
View details for PubMedID 34673889
Immunoprofiling of Nonarteritic Anterior Ischemic Optic Neuropathy.
Translational vision science & technology
2021; 10 (8): 17
Purpose: Nonarteritic anterior ischemic optic neuropathy (NAION) is a common acute optic neuropathy in those older than 50 years. There is no blood diagnostic test or efficient treatment for NAION. We investigated the suitability of blood inflammatory proteins as biomarkers and therapeutic targets of NAION.Methods: We conducted an exploratory, cross-sectional case-control study including 18 patients with NAION (n = 5 acute, and n = 13 chronic) and 9 controls. NAION was confirmed by clinical examination and optical coherence tomography. Subjects underwent peripheral blood collection; plasma was isolated within 2 hours and analyzed using a 76-plex array of cytokines, chemokines, and growth factors.Results: In acute NAION, there was increased peripapillary retinal thickness on optical coherence tomography consistent with optic disc edema. Plasma profiling revealed dramatic changes in inflammatory proteins in NAION. Statistical analysis generated a list of 20 top-ranked molecules in NAION, with 15% overlap in acute and chronic NAION. Principal component analysis, hierarchical clustering, and Spearman correlation generally segregated controls, acute and chronic NAION, with some overlap. Longitudinal data from one patient demonstrated an evolving inflammatory pattern from acute to chronic NAION. In acute NAION, Eotaxin-3, MCP-2, TPO, and TRAIL were the top biomarker candidates. In chronic NAION, IL-1alpha and CXCL10 emerged as the strongest therapeutic targets.Conclusions: Post-NAION inflammation occurs in both acute and chronic NAION. Statistical analysis of plasma profile changes generated a list of 20 potential biomarker and therapeutic targets of NAION.Translational Relevance: We identified blood molecular targets to improve NAION diagnosis and treatment.
View details for DOI 10.1167/tvst.10.8.17
View details for PubMedID 34264294
Topographic Quadrant Analysis of Peripapillary Superficial Microvasculature in Optic Disc Drusen
FRONTIERS IN NEUROLOGY
2021; 12: 666359
Background: Limited information is known about the topographic effect of optic disc drusen (ODD) on peripapillary retinal nerve fibers and microvasculature. Objective: This study aims to understand the structural and functional impact of ODD in different quadrants of the optic disc. Methods: We performed a retrospective case-control study of 22 ODD patients (34 eyes) and 26 controls (33 eyes) to compare optical coherence tomography (OCT) retinal nerve fiber layer (RNFL), OCT angiography (OCTA), and corresponding static perimetry mean deviation (MD) calculated using the modified Garway-Heath map in different quadrants of the optic disc. OCTA was analyzed using custom MATLAB script to measure six parameters in a peripapillary annulus with large vessel removal: vessel area density (VAD), vessel skeleton density (VSD), vessel perimeter index (VPI), vessel complexity index (VCI), flux, and vessel diameter index (VDI). Results: Quadrant analysis revealed that OCTA VAD and VCI were significantly decreased in superior, nasal, and inferior but not temporal quadrant. RNFL, VSD, and VPI were significantly impacted only in the superior and nasal quadrants. Corresponding visual field MDs in all ODD eyes were not different in the four quadrants, although eyes with MD equal or worse than -5 dB (32%) had worst visual field corresponding to the superior quadrant of the optic disc (inferior arcuate visual field). Structure-structure comparison of OCT and OCTA showed high correlation of RNFL with multiple OCTA measurements in the superior, nasal, and inferior quadrants but not temporal quadrant. Structure-function analysis revealed significant correlation of VAD and VCI and visual field MD in every quadrant, but RNFL was only significantly correlated in the superior and inferior quadrants. Conclusions: Peripapillary VAD and VCI are decreased in more quadrants than RNFL, supporting the clinical utility of performing OCTA in addition to OCT. Consistent with the most common locations of ODD, five OCT/OCTA measurements (VAD, VCI, RNFL, VSD, VPI) are decreased in the superior and nasal quadrants. OCT/OCTA measurements were significantly impacted in contrast to the relatively mild effect on corresponding visual field MD, consistent with the idea that a decrease in objective structural and vascular measurements occurs without parallel change in subjective visual function in ODD.
View details for DOI 10.3389/fneur.2021.666359
View details for Web of Science ID 000656846300001
View details for PubMedID 34093412
View details for PubMedCentralID PMC8170317
Hypoxia-induced inflammation: Profiling the first 24-hour posthypoxic plasma and central nervous system changes.
2021; 16 (3): e0246681
Central nervous system and visual dysfunction is an unfortunate consequence of systemic hypoxia in the setting of cardiopulmonary disease, including infection with SARS-CoV-2, high-altitude cerebral edema and retinopathy and other conditions. Hypoxia-induced inflammatory signaling may lead to retinal inflammation, gliosis and visual disturbances. We investigated the consequences of systemic hypoxia using serial retinal optical coherence tomography and by assessing the earliest changes within 24h after hypoxia by measuring a proteomics panel of 39 cytokines, chemokines and growth factors in the plasma and retina, as well as using retinal histology. We induced severe systemic hypoxia in adult C57BL/6 mice using a hypoxia chamber (10% O2) for 1 week and rapidly assessed measurements within 1h compared with 18h after hypoxia. Optical coherence tomography revealed retinal tissue edema at 18h after hypoxia. Hierarchical clustering of plasma and retinal immune molecules revealed obvious segregation of the 1h posthypoxia group away from that of controls. One hour after hypoxia, there were 10 significantly increased molecules in plasma and 4 in retina. Interleukin-1β and vascular endothelial growth factor were increased in both tissues. Concomitantly, there was significantly increased aquaporin-4, decreased Kir4.1, and increased gliosis in retinal histology. In summary, the immediate posthypoxic period is characterized by molecular changes consistent with systemic and retinal inflammation and retinal glial changes important in water transport, leading to tissue edema. This posthypoxic inflammation rapidly improves within 24h, consistent with the typically mild and transient visual disturbance in hypoxia, such as in high-altitude retinopathy. Given hypoxia increases risk of vision loss, more studies in at-risk patients, such as plasma immune profiling and in vivo retinal imaging, are needed in order to identify novel diagnostic or prognostic biomarkers of visual impairment in systemic hypoxia.
View details for DOI 10.1371/journal.pone.0246681
View details for PubMedID 33661927
Phase 1b randomized controlled study of short course topical recombinant human nerve growth factor (rhNGF) for neuroenhancement in glaucoma: safety, tolerability and efficacy measure outcomes.
American journal of ophthalmology
No approved therapies directly target retinal ganglion cells (RGCs) for neuroprotection or neuroenhancement in glaucoma. Recombinant human nerve growth factor (rhNGF) has been shown to promote RGC survival and function in animal models of optic neuropathy. Here we evaluate safety, tolerability, and efficacy of short-term, high-dose rhNGF eye drops versus placebo in a cohort of glaucoma patients.This study is a single-center, randomized, double-masked, vehicle-controlled, parallel group study designed to assess safety and tolerability as well as short-term neuroenhancement of structure and function (Clinicaltrials.gov NCT02855450). Sixty open-angle glaucoma patients were randomized 40:20 to receive either 180 μg/ml rhNGF or vehicle control eye drops in both eyes, three times daily for 8 weeks, with a 24-week post-treatment follow-up. One eye was officially selected as the study eye, although both eyes were studied and dosed. Primary endpoints were safety, as assessed through adverse events, and tolerability, as assessed through patient reported outcomes. Secondary outcome measures included best corrected visual acuity (BCVA), Humphrey visual field (HVF), electroretinogram (ERG), and optical coherence tomography (OCT) of retinal nerve fiber layer (RNFL) thickness at baseline, after 8 weeks of treatment, and at 4 and 24 weeks after treatment (12- and 32-weeks total).Of the 60 randomized subjects, 23 were female (38%) and the average age was 66.1 years. Through week 32, there were no treatment-related serious adverse events, including no unexpectedly severe progression of optic neuropathy, no adverse events affecting ocular function or pressure, and no drug-related systemic toxicity. Topical high-dose rhNGF was tolerated well, with low level of symptom burden mainly eliciting periocular ache (in 52% of treated, 5% of placebo) and only 3 patients (7.5%) discontinuing treatment due to discomfort, out of whom 1 patient (2.5%) prematurely withdrawing from the study. There were no statistically significant differences in global indices of HVF, and no meaningful differences in total, quadrant, or clock-hour mean RNFL thickness between the groups, although both of these function and structure measures showed non-significant trends towards significance in favor of rhNGF. Real-world participant data was used to generate an estimate of cohort size needed to power subsequent studies.rhNGF is safe and tolerable in a topical 180 μg/ml formulation. Although no statistically significant short-term neuroenhancement was detected in this trial, given the strong effects of NGF in preclinical models and trends detected in this study, analysis for efficacy in a neuroprotection trial is warranted.
View details for DOI 10.1016/j.ajo.2021.11.002
View details for PubMedID 34780798
Corneal Light Scatter After Ultrathin Descemet Stripping Automated Endothelial Keratoplasty Versus Descemet Membrane Endothelial Keratoplasty in Descemet Endothelial Thickness Comparison Trial: A Randomized Controlled Trial.
PURPOSE: To compare the degree of corneal light scatter as measured by densitometry in ultrathin Descemet Stripping Automated Endothelial Keratoplasty (UT-DSAEK) and Descemet Membrane Endothelial Keratoplasty (DMEK) in the Descemet endothelial thickness comparison trial.METHODS: This was a prespecified secondary analysis of the Descemet endothelial thickness comparison trial, which was a prospective, randomized controlled trial. Subjects with isolated endothelial dysfunction were enrolled and were randomized to either UT-DSAEK or DMEK. Corneal opacity was quantitatively measured by Pentacam densitometry (OCULUS) at 3, 6, and 12 months.RESULTS: Fifty eyes of 38 patients were enrolled at the Casey Eye Institute at Oregon Health & Science University and the Byers Eye Institute at Stanford University. Corneal densitometry for the anterior and posterior layers improved in both UT-DSAEK and DMEK after surgery. The decrease was more pronounced in the posterior layer for both groups. However, there was no difference in the degree of corneal light scatter between UT-DSAEK and DMEK at postoperative month 12, and no difference in change in densitometry was observed between the 2 arms from baseline to month 12.CONCLUSIONS: Both UT-DSAEK and DMEK experience an improvement in the degree of corneal light scatter after surgery. However, there was no difference in densitometry between the 2 groups at month 12. Therefore, other factors such as higher order aberrations in the posterior cornea rather than stromal-stromal interface haze mediate the superior visual outcomes in DMEK compared with UT-DSAEK.
View details for DOI 10.1097/ICO.0000000000002256
View details for PubMedID 31939923
Vision loss in optic disc drusen correlates with increased macular vessel diameter and flux and reduced peripapillary vascular density.
American journal of ophthalmology
To determine the key optical coherence tomography (OCT) and OCT angiography (OCTA) parameters that correlate with visual field loss in optic disc drusen (ODD).Retrospective cross-sectional study..Single academic center.17 patients with ODD (29 eyes) and 35 age-matched controls (53 eyes) INTERVENTION OR OBSERVATION PROCEDURES: Static perimetry, OCT and OCTA imaging of optic disc and macula.static perimetry, OCT, and OCTA measurements.We investigated the relationship between static perimetry and 14 OCT/OCTA measurements in patients with ODD vs. age-matched controls and found 5 key measurements that most correlated with visual field loss included: peripapillary retinal nerve fiber layer (RNFL), macular ganglion cell complex (GCC), peripapillary vessel area density (VAD), macular vessel diameter (VD) and flux. Hierarchical clustering of these 5 measurements vs. all clinical characteristics revealed 3 distinct clusters. ODD and control eyes with no visual field loss (mean deviation (MD) > -2.0 dB) had high RNFL, GCC, and low macular VD and flux. ODD eyes with mild visual field loss (MD -2.0 to -5.0 dB) had high RNFL, GCC, and increased macular VD and flux. ODD eyes with moderate/severe visual field loss (MD < -5.0 dB) had decreased RNFL, GCC, peripapillary VAD, and increased macular VD and flux.OCT and OCTA provided objective measurements that can help predict visual field loss in ODD. Our data suggest that increased macular flow may be an early biomarker of visual field loss in ODD, while decreased peripapillary vessel density and RNFL thickness are late biomarkers of visual field loss in ODD.
View details for DOI 10.1016/j.ajo.2020.04.019
View details for PubMedID 32360344
Exome sequencing of Finnish isolates enhances rare-variant association power.
Exome-sequencing studies have generally been underpowered to identify deleterious alleles with a large effect on complex traits as such alleles are mostly rare. Because the population of northern and eastern Finland has expanded considerably and in isolation following a series of bottlenecks, individuals of these populations have numerous deleterious alleles at a relatively high frequency. Here, using exomesequencing of nearly 20,000 individuals from these regions, we investigate the role of rare coding variants in clinically relevant quantitative cardiometabolic traits. Exome-wide association studies for 64 quantitative traits identified 26 newly associated deleterious alleles. Of these 26 alleles, 19 are either unique to or more than 20 times more frequent in Finnish individuals than in other Europeans and show geographical clustering comparable to Mendelian disease mutations that are characteristic of the Finnish population. We estimate that sequencing studies of populations without this unique history would require hundreds of thousands to millions of participants to achieve comparable association power.
View details for DOI 10.1038/s41586-019-1457-z
View details for PubMedID 31367044
Multiregion Quantification of Extracellular Signal-regulated Kinase Activity in Renal Cell Carcinoma.
European urology oncology
To personalize treatment for renal cell carcinoma (RCC), it would be ideal to confirm the activity of druggable protein pathways within individual tumors. We have developed a high-resolution nanoimmunoassay (NIA) to measure protein activity with high precision in scant specimens (eg, fine needle aspirates [FNAs]). Here, we used NIA to determine whether protein activation varied in different regions of RCC tumors. Since most RCC therapies target angiogenesis by inhibiting the vascular endothelial growth factor (VEGF) receptor, we quantified phosphorylation of extracellular signal-regulated kinase (ERK), a downstream effector of the VEGF signaling pathway. In 90 ex vivo FNA biopsies sampled from multiple regions of 38 primary clear cell RCC tumors, ERK phosphorylation differed among patients. In contrast, within individual patients, we found limited intratumoral heterogeneity of ERK phosphorylation. Our results suggest that measuring ERK in a single FNA may be representative of ERK activity in different regions of the same tumor. As diagnostic and therapeutic protein biomarkers are being sought, NIA measurements of protein signaling may increase the clinical utility of renal mass biopsy and allow for the application of precision oncology for patients with localized and advanced RCC. PATIENT SUMMARY: In this report, we applied a new approach to measure the activity of extracellular signal-regulated kinase (ERK), a key cancer signaling protein, in different areas within kidney cancers. We found that ERK activity varied between patients, but that different regions within individual kidney tumors showed similar ERK activity. This suggests that a single biopsy of renal cell carcinoma may be sufficient to measure protein signaling activity to aid in precision oncology approaches.
View details for DOI 10.1016/j.euo.2018.09.011
View details for PubMedID 31412000
Exposure to NO2, CO, and PM2.5 is linked to regional DNA methylation differences in asthma
2018; 10: 2
DNA methylation of CpG sites on genetic loci has been linked to increased risk of asthma in children exposed to elevated ambient air pollutants (AAPs). Further identification of specific CpG sites and the pollutants that are associated with methylation of these CpG sites in immune cells could impact our understanding of asthma pathophysiology. In this study, we sought to identify some CpG sites in specific genes that could be associated with asthma regulation (Foxp3 and IL10) and to identify the different AAPs for which exposure prior to the blood draw is linked to methylation levels at these sites. We recruited subjects from Fresno, California, an area known for high levels of AAPs. Blood samples and responses to questionnaires were obtained (n = 188), and in a subset of subjects (n = 33), repeat samples were collected 2 years later. Average measures of AAPs were obtained for 1, 15, 30, 90, 180, and 365 days prior to each blood draw to estimate the short-term vs. long-term effects of the AAP exposures.Asthma was significantly associated with higher differentially methylated regions (DMRs) of the Foxp3 promoter region (p = 0.030) and the IL10 intronic region (p = 0.026). Additionally, at the 90-day time period (90 days prior to the blood draw), Foxp3 methylation was positively associated with NO2, CO, and PM2.5 exposures (p = 0.001, p = 0.001, and p = 0.012, respectively). In the subset of subjects retested 2 years later (n = 33), a positive association between AAP exposure and methylation was sustained. There was also a negative correlation between the average Foxp3 methylation of the promoter region and activated Treg levels (p = 0.039) and a positive correlation between the average IL10 methylation of region 3 of intron 4 and IL10 cytokine expression (p = 0.030).Short-term and long-term exposures to high levels of CO, NO2, and PM2.5 were associated with alterations in differentially methylated regions of Foxp3. IL10 methylation showed a similar trend. For any given individual, these changes tend to be sustained over time. In addition, asthma was associated with higher differentially methylated regions of Foxp3 and IL10.
View details for PubMedID 29317916
Genetic Variant Selection: Learning Across Traits and Sites
2016; 202 (2): 439-?
We consider resequencing studies of associated loci and the problem of prioritizing sequence variants for functional follow-up. Working within the multivariate linear regression framework helps us to account for the joint effects of multiple genes; and adopting a Bayesian approach leads to posterior probabilities that coherently incorporate all information about the variants' function. We describe two novel prior distributions that facilitate learning the role of each variable site by borrowing evidence across phenotypes and across mutations in the same gene. We illustrate their potential advantages with simulations and reanalyzing a data set of sequencing variants.
View details for DOI 10.1534/genetics.115.184572
View details for Web of Science ID 000371304600010
View details for PubMedID 26680660
View details for PubMedCentralID PMC4788227