Publications

Background: Medical big data analytics has revolutionized the human healthcare system by introducing processes that facilitate rationale clinical decision making, predictive or prognostic modelling of the disease progression and management, disease surveillance, overall impact on public health and research. Although, the electronic medical records (EMR) system is the digital storehouse of rich medical data of a large patient cohort collected over many years, the data lack sufficient structure to be of clinical value for applying deep learning methods and advanced analytics to improve disease management at an individual patient level or for the discipline in general. Ophthatome™ captures data contained in retrospective electronic medical records between September 2012 and January 2018 to facilitate translational vision research through a knowledgebase of ophthalmic diseases.

Methods: The electronic medical records data from Narayana Nethralaya ophthalmic hospital recorded in the MS-SQL database was mapped and programmatically transferred to MySQL. The captured data was manually curated to preserve data integrity and accuracy. The data was stored in MySQL database management system for ease of visualization, advanced search functions and other knowledgebase applications.

Results: Ophthatome™ is a comprehensive and accurate knowledgebase of ophthalmic diseases containing curated clinical, treatment and imaging data of 581,466 ophthalmic subjects from the Indian population, recorded between September 2012 and January 2018. Ophthatome™ provides filters and Boolean searches with operators and modifiers that allow selection of specific cohorts covering 524 distinct ophthalmic disease types and 1800 disease sub-types across 35 different anatomical regions of the eye. The availability of longitudinal data for about 300,000 subjects provides additional opportunity to perform clinical research on disease progression and management including drug responses and management outcomes. The knowledgebase captures ophthalmic diseases in a genetically diverse population providing opportunity to study genetic and environmental factors contributing to or influencing ophthalmic diseases.

Conclusion: Ophthatome™ will accelerate clinical, genomic, pharmacogenomic and advanced translational research in ophthalmology and vision sciences.

Abstract
Classical homocystinuria is the most common cause of isolated homocystinuria. The variants of the CBS gene remain unidentified in Indian children with this disorder. Based on the hallmark clinical features, family history, and/or biochemical clues for classical homocystinuria, 16 children below the age of 18 years were evaluated by Sanger sequencing of the coding exons of CBS gene with flanking intronic regions. The common C677T variant of the MTHFR gene was also screened by restriction fragment length polymorphism. Fifteen children were clinically suspected of having classical homocystinuria and one asymptomatic child with positive family history. Only seven children had biochemical features of classical homocystinuria. Sanger sequencing of the CBS gene confirmed 15 different pathogenic or likely pathogenic variants in 14 cases. Of these, seven variants were novel (three frameshift deletions, two nonsense, one missense, one splice site variant) and were predicted to be deleterious by Mutation Taster software. Seven cases were homozygous, another six were compound heterozygous, and one case was single heterozygous in the study. None of the three most frequent mutations reported worldwide viz., I278T, G307S, and IVS 11-2A>C were found in our cohort. No variants were detected in the exons 2, 8, 12, and 14 as compared to reported literature. Eleven out of 15 variants were associated with the conserved catalytic domain of the CBS polypeptide. The MTHFR polymorphism C677T was observed in heterozygous state in six cases. Our study reports the detailed genotype and seven novel variants in the CBS gene, causing classical homocystinuria in Indian children. The genetic analysis will help to offer accurate genetic counseling, prenatal diagnosis, and development of mutation-based novel therapeutic strategies.

Abstract This study explores the etiology and lead time to treatment for infantile spasm (IS) patients and their effect on treatment responsiveness, in a limited resource setting. Patients with IS onset age ≤12 months’, seen over 3 years were recruited retrospectively. Clinical information, neuroimaging and genetic results retrieved. Patients categorized into three primary etiological groups: Structural (including Structural Genetic), Genetic, and Unknown. The effect of etiology and lead time from IS onset to initiating appropriate treatment on spasm resolution, evaluated. Total 113 patients were eligible. Mean IS onset age was 6.86(±4.25) months (M: F 3.3:1). Patients were grouped into: Structural 85, Genetic 11 and Unknown 17. Etiology was ascertained in 94/113 (83.1%) with neonatal hypoglycemic brain injury (NHBI) being the most common (40/113, 36%). A genetic etiology identified in 17 (including 6 Structural Genetic, of which five had Tuberous Sclerosis). Structural group was less likely to be treatment resistant (p = 0.013, OR 0.30 [0.12-0.76]). Median treatment lead time – 60 days. Longer lead time to treatment was significantly associated with resistant spasms (χ2 for trend = 10.0, p = 0.0015). NHBI was the commonest underlying cause of IS. There was significant time lag to initiating appropriate treatment, affecting treatment responsiveness.

Keywords: ataxia, CJD, familial, vocal cord palsy, autosomal dominant.

BACKGROUND Genome-wide polygenic scores (GPS) integrate information from many common DNA variants into a single number. Because rates of coronary artery disease (CAD) are substantially higher among South Asians, a GPS to identify high-risk individuals may be particularly useful in this population.
OBJECTIVES This analysis used summary statistics from a prior genome-wide association study to derive a new GPSCAD for South Asians.
METHODS This GPSCAD was validated in 7,244 South Asian UK Biobank participants and tested in 491 individuals from a case-control study in Bangladesh. Next, a static ancestry and GPSCAD reference distribution was built using wholegenome sequencing from 1,522 Indian individuals, and a framework was tested for projecting individuals onto this static ancestry and GPSCAD reference distribution using 1,800 CAD cases and 1,163 control subjects newly recruited in India.
RESULTS The GPSCAD, containing 6,630,150 common DNA variants, had an odds ratio (OR) per SD of 1.58 in South Asian UK Biobank participants and 1.60 in the Bangladeshi study (p < 0.001 for each). Next, individuals of the Indian casecontrol study were projected onto static reference distributions, observing an OR/SD of 1.66 (p < 0.001). Compared with the middle quintile, risk for CAD was most pronounced for those in the top 5% of the GPSCAD distribution—ORs of 4.16, 2.46, and 3.22 in the South Asian UK Biobank, Bangladeshi, and Indian studies, respectively (p < 0.05 for each). CONCLUSIONS The new GPSCAD has been developed and tested using 3 distinct South Asian studies, and provides a generalizable framework for ancestry-specific GPS assessment. (J Am Coll Cardiol 2020;76:703–14) © 2020 by the American College of Cardiology Foundation.

KMT2B-related dystonia is characterized by childhood-onset and progressive disease course with prominent lower-limb, cervical, cranial, and laryngeal involvement [ 1 ]. Developmental delay and intellectual disability are commonly reported [ 1 , 2 ]. Here we are reporting an Indian patient with KMT2B-related dystonia who had a disease onset at the age of 19 years. Other interesting features in our patient were normal cognition, dystonic opisthotonus, and absence of oromandibular and laryngeal dystonia.

We report a family with a spectrum of short stature, craniofacial dysmorphism, and digital anomalies in a father and 2 daughters, with the youngest (proband) displaying a severe phenotype. Clinically, autosomal dominant Robinow syndrome (ADRS) was diagnosed. Whole-exome sequencing identified a heterozygous pathogenic BMP2 variant in the father and his daughters. The phenotype of short stature, facial dysmorphism, and skeletal anomalies with or without cardiac anomalies related to BMP2 haploinsufficiency has some facial and digital resemblance to ADRS. Although this variant segregated in the affected members, it failed to explain the severe phenotype of the proband. A reanalysis of the girl’s raw data confirmed 2 disorders: a de novo likely pathogenic DVL1 variant implicated in ADRS and the familial BMP2 variant. A close interplay of high-throughput sequencing and deep phenotyping unraveled the complexities of the blended phenotype in the proband.

Cervical cancer is a growing and serious problem world-wide in women, but more acute in developing countries especially in Indian subcontinent. The main causative agent for the disease is Human Papilloma Virus (HPV). The history of the cervical cancer goes back to eighteenth century as the HPV infection is reported since 1800s. Presently, the genetic structure of HPV is well defined. Several screening tests including cytology and visual based screening and high risk HPV testing are available. Also available are various clinical and commercial diagnostic tests. However due to the lack of awareness and population-based screening programs, the morbidity and mortality rate is alarmingly high. There are new emerging biomarkers including E6/E7 mRNA, p16ink4a, markers of aberrant S-phase induction, chromosomal abnormalities and miRNAs along with advanced genotyping methods. These markers have clinical significance and are helpful in disease prevention and management. Further, recent advancement in the field of metagenomics has increased the prospects of identifying newer microbes, viruses hitherto reported thus far in the context of HPV infection. Analysis of HPV cases using modern tools including genotyping using more powerful biomarkers is envisaged to enhance the prospects of early diagnosis, better prognosis, more reliable treatment and eventual management of the disease.

The T-box4 (TBX4) gene (OMIM *601719) belongs to the T-box family of transcription regulators that share a conserved homology domain and are expressed at specific sites during various stages of embryonic development. Tbx4 has been found to be a crucial transcriptional regulator in embryonic hindlimb development in animal models. Monoallelic variants in the TBX4 gene are reported to be associated with skeletal defects of the pelvis and lower limbs. We report here a fetus with a novel multiple malformation syndrome associated with sacrococcygeal agenesis, bilateral lower limb aplasia, hypoplastic left heart, bilateral lung hypoplasia, hydroureteronephrosis, and nonimmune fetal hydrops, found to have a homozygous nonsense variant in the TBX4 gene. We propose that biallelic variants in the TBX4 gene are associated with a severe syndromic phenotype of sacrococcygeal agenesis and lower limb reduction defects.

Abstract
Background
Studies evaluating next‐generation sequencing (NGS) for retinal disorders may not reflect clinical practice. We report results of retrospective analysis of patients referred for clinical testing at two institutions (US and India).

Methods
This retrospective study of 131 patients who underwent clinically validated targeted NGS or exome sequencing for a wide variety of clinical phenotypes categorized results into a definitive, indeterminate, or negative molecular diagnosis.

Results
A definitive molecular diagnosis (52%) was more common in the India cohort (62% vs. 39%, p = .009), while an indeterminate molecular diagnosis occurred only in the US cohort (12%). In the US cohort, a lower diagnostic rate in Hispanic, non‐Caucasians (23%) was seen compared to Caucasians (57%). The India cohort had a high rate of homozygous variants (61%) and different frequency of genes involved compared to the US cohort.

Conclusion
Despite inherent limitations in clinical testing, the diagnostic rate across the two cohorts (52%) was similar to the 50%–65% diagnostic rate in the literature. However, the diagnostic rate was lower in the US cohort and appears partly explained by racial background. The high rate of consanguinity in the Indian population is reflected in the high rate of homozygosity for pathogenic mutations and may have implications for population level screening and genetic counseling. Clinical laboratories may note diagnostic rates that differ from the literature, due to factors such as heterogeneity in racial background or consanguinity rates in the populations being tested. This information may be useful for post‐test counseling.