The 2021 annual conference hosted by the American Society of Human Genetics was packed full of revolutionary new findings across a wide range of areas in the field of genomics. The implementation of genome sequencing in clinical settings is rapidly growing, with leading researchers constantly finding new ways to solve adverse clinical and diagnostic problems using the newest sequencing technologies. Merogenomics has been given the honor of being invited to cover the conference, and we hope to showcase some of the major themes and highlights of the event . Throughout this conference, there was an abundance of opportunities to learn about the current state of medical genomics, the focus of novel research in the field, and some of the major barriers that must be overcome before seeing an increase in accessibility and overall implementation of these incredible new technologies. A recurring point that continues to be emphasized is that the field of medicine is working hard to keep up with advances in genome sequencing technologies but remains far behind due to difficulty in implementing some of the new, unfamiliar, and expensive protocols. Here we have focused on recent research and developments in clinical genomics, intending to highlight some of the progress being made towards optimization and large-scale implementation of genome sequencing in clinical settings.

Long-read sequencing

Although implementing next-generation sequencing (NGS) has opened possibilities in precision medicine, diagnostic rates for clinical sequencing are only ~35%. Continual research has demonstrated the technology’s shortcomings, highlighting strategies to further advance diagnostic rates.

Among the many methods to improve diagnostic rate, long-read sequencing (LRS) is a promising tool to help overcome the limitations of the short-read in NGS. As the name suggests, they’re just longer. When typical short-read sequences (SRS) span 100-300 base pairs (bp) in length, LRS clearly demonstrates its power by being capable of spanning a whole megabase (Mb), which is the equivalent of 1,000,000 bp. This can cover at least 15% of the genome that SRS does not capture well or at all. In addition, LRS offers a plethora of advantages such as improved calling in NGS dead zones, lower validation costs, lower false positives, superior structural variant detection, and analysis of methylation data. Although not yet clinically validated due to its preliminary stages of application, the future looks promising. However, there is not enough LRS data available yet to allow for fully powered validation studies. Still, LRS has even demonstrated a high concordance with short-read sequence calling for small variants which supports its utility.

From the Seattle Children’s hospital, applying his research from the genome sciences and pediatrics departments, Dr. Danny Miller uses targeted-long-read sequencing (T-LRS) to help identify disease-causing variation. Through the use of either cas9 mediated targeting or PCR enrichment, T-LRS can target a 15Mb region of DNA. In combination with adaptive sampling to allow sequencing in real-time, T-LRS was found to achieve 20-50x coverage when identifying an X-chromosome inversion in Lesch-Nyhan syndrome. This finding helped resolve triplet repeat expansions and complex copy number expansions and their corresponding methylation states, helping provide further clinical relevance. Another benefit of selecting targeted LRS is that it lowers the per-sample cost of sequencing selected genes or regions to a price point compared to short-read whole-genome sequencing.

Another exciting application of LRS technology was detailed by Dr. Mathew Bainbridge, who specializes in translating clinical genomics in precision medicine at the Rady’s children's hospital. Achieving a coverage of 10-30x with LRS, they established a diagnosis for epilepsy, syndromic autism in previously undiagnosed cases. The LRS was further used to gather insights on complex undermined syndromes and sudden infant death. Dr. Bainbridge presented recent findings regarding a novel frameshift mutation in IKBKG in one of four undiagnosed immune phenotype cases. IKBKG is a gene that is known to cause immune deficiencies when mutated and is categorized as “likely pathogenic” in ClinVar in the NCBI database. With furthering the number of novel findings that could provide clinically actionable decisions, Dr. Bainbridge suggests that LRS could be the scientific method of choice to eliminate unnecessary diagnostic tests.

From the department of human health sciences at Kyoto University, Ph.D. student Kiyose Hiroki explained how his research endeavors focused on using the Oxford Nanopore long-read RNA sequencer to examine differences in the transcriptome of healthy liver cells in comparison to hepatocellular carcinoma (HCC) cells. By examining the transcriptome abnormalities between healthy and cancerous tissues, one can discover new cancer-driving mutations and mechanisms that lead to further discoveries and strategies to combat cancer. In this case, the long-reads detected approximately 15.7% novel transcripts that were not yet seen in registered databases. Of the findings, a highly conserved transcript (MYT1L) was demonstrated to have tumor suppressor-like effects in the liver. These findings expand the previously reported observations, which describe MYT1L as a neuron-specific factor playing a role in neuronal differentiation and maintenance. Further, HCC and liver transcriptomics comparisons also presented a transposable element derived transcript from oncogene MET (L1-MET). MET was observed to increase proliferation and support tumor growth in HCC.

Expansion of known variant database

A major drawback faced in the healthcare system when approached with rare undiagnosed conditions is a lack of background information on the condition, a generally poor understanding of the cause, and a shortage of available treatment options. When it comes to genetic disorders, this can commonly be attributed to the absence of data surrounding genetic associations of the disease. Genetic variants and pedigree profiles associated with these conditions are typically unknown, and as a result many healthcare practitioners faced with these cases may be unfamiliar with the conditions and/or possible treatment options for these patients. A recurring topic of focus that was highlighted in several talks throughout the ASHG conference was the importance of identifying rare genetic variants that are widely unknown, especially in regions of the genome that have been difficult to analyze with the most commonly implemented sequencing technologies.

Samuel Peterson from the Oregon National Primate Research Center (ONPRC) gave an exciting talk highlighting their recent work in establishing the Macaque Genotype and Phenotype Resource (mGAP). Samuel’s group has been collecting a broad spectrum of valuable data from a large Macaque cohort housed at the ONPRC. The database includes whole genome and whole exome sequence data, phenotype data, extensive health records, and pedigree information for all the Rhesus Macaques at their facility. This catalog of invaluable data presents great opportunities for fellow researchers to identify ideal animal models with detected pathogenic or likely pathogenic variants and a depth of phenotype data to go along with it. Establishing animal models that are homologous to some of the human cases faced by other researchers is extremely useful and made easier with the abundant availability of extensive pedigree data, allowing researchers to understand inheritance patterns of their disease of interest. This is an incredible resource and has great potential to help improve understanding of rare genetic conditions in humans.

Another fascinating talk by Wouter Steyaert from Radboud University Nijmegen Medical Center was focused on identifying clinically relevant variants, specifically in homologous regions. Regions of the genome that have high sequence homology with other regions can be difficult to analyze with current sequencing and bioinformatics protocols. Steyaert’s group believed that identifying rare and unknown genetic variants in these regions could have drastic implications for increasing diagnostic rates for genetic conditions in undiagnosed individuals. In his talk, Steyaert presented their novel Cameleolyser algorithm that was developed to detect rare variants in these homologous regions with improved effectiveness. Their deployment of the Cameleolyser algorithm was used to identify rare homozygous deletions and rare single nucleotide variants in these challenging regions. The detected variants included ultra-rare homozygous deletions in well known OMIM genes, and loss-of-function variants, many of which were linked to gene conversions. Additionally, their analysis of sequence data from patients with hearing impairment was able to identify homozygous loss of function variants and deletions in STRC, a gene widely associated with deafness. This remarkable technology has great potential for identification of rare variants in genetic regions that have previously been difficult to analyze.

Alden Yen-Wen Huang from the California Center for Rare Diseases at UCLA had an interesting approach for improving diagnostic yields of RNA sequencing (RNA-Seq) technology. Alden’s work at UCLA was focused on a CRISPR Cas9 based system for depletion of highly abundant transcripts to help improve coverage of lesser expressed transcripts. In the most common RNA-Seq protocols currently used, whole-transcriptome analyses predominantly reveal commonly known variants in the most highly transcribed genes. Although a significant proportion of known disease-associated genes are observed in RNA-Seq, many of these are expressed at levels that are too low for comprehensive analysis. Alden’s group developed a novel method that uses CRISPRclean, a Cas9 based RNA depletion protocol with guide RNAs to specifically remove targeted RNA-Seq library fragments from targeted genes, allowing for a significant increase in coverage of the remaining untargeted genes. This method allows for increased detection of rare variants that are expressed in lower quantities in whole transcriptome analysis, which gives an opportunity to expand the proportion of the genome that is interpretable by RNA-Seq. This technology can drastically help with increasing the overall implementation of RNA-Seq protocols as more rare variants can be detected, which can eventually increase diagnostic rates of rare variants that were previously unknown or difficult to analyze due to their low abundance in RNA-Seq libraries.

Using automation to improve diagnostic rates and streamline genetic testing protocols

The rapid advancement of genetic testing technology in recent years has created a disconnect between the state of genomics research and the accessibility and implementation of genome based diagnostics and therapies being utilized at the clinical level. Many proponents of genome sequencing would argue that genome based protocols have great potential but still remain underutilized in clinical settings. The lack of implementation can largely be attributed to the unfamiliarity and overall absence of exposure that most first line physicians face. An exciting theme in the conference was the use of automation and AI based technologies to streamline genetic testing protocols and help overcome the barrier of physician unfamiliarity. Automating the analysis of diagnostic genetic testing results also has great potential to drastically reduce the time to diagnosis and could even help reduce costs for the healthcare system over time.

Dr. Stephen Francis Kingsmore from the Rady Children’s Institute for Genomic Medicine gave an exciting talk about a new virtual automated system for diagnosing and managing patients based on genetic testing results. On top of that, this software boasts an incredible turnover rate, with an ability to give results in only 13.5 hours. The software, known as GTRx, was able to establish a diagnosis in 13.5 hours by improving the efficiency of whole genome sequencing and informatic analysis, while also interpreting health records and their associations with genetic sequencing data. To demonstrate the effectiveness of the program, they used it to identify 563 severe genetic diseases, and establish treatment options. A large portion of these diagnoses were then validated by a group of five clinical geneticists, who agreed upon 189 of the first 190 treatment suggestions proposed by the program. GTRx was able to detect many different types of variants, including single nucleotide variants, insertion-deletion, structural variants, and copy number variants. GTRx and similar technologies focused around automating the analysis of genetic testing results have great potential for optimizing the acute management of children with rapidly progressing genetic diseases and could drastically improve clinical outcomes for these patients.

From the department of cardiovascular medicine at mayo clinic, Alexandra Miller gave insights to how automation techniques can benefit the diagnostic rate of familial hypercholesterolemia (FH). Increasing the diagnostic rate of FH is of utmost importance, considering that 90% of individuals go undiagnosed. Diagnostic criteria were based on three metrics: 1) male above 18 years of age 2) LDL concentration >190 mg/dL and 3) an upcoming appointment. Once meeting these criteria, the research coordinator analyzed electronic health records (EHR) to investigate secondary causes of FH like hypothyroidism, cholestatic liver disease, and nephrotic syndrome. After going through EHRs to establish a dutch lipid clinic (DLCN) score to qualify those for sequencing, patients were mailed informed consent forms along with a sample collection kit.

Using a targeted gene panel, LDLR, APOB, PCSK9 genes were investigated. Mutational distribution was highest in LDLR, followed by APOB and PCSK9, with the single nucleotide variants (SNVs) being the most frequent mutation type. Interestingly, the average LDL concentration was highest in patients possessing mutations other than SNVs such as copy number variations (CNVs), insertions and deletions, duplications, and splice site variants. Of the total 84 individuals sequenced, 45% were shown to possess a pathogenic/likely pathogenic (P/LP) variant. These P/LP findings present an opportunity for those individuals to seek genetic counseling and initiate downstream cascade testing within the family. This novel, two-step protocol for identifying individuals with monogenic FH will be a focal point for the Mayo clinic in the coming future.

Another talk by Deanna Brockman from Massachusetts general hospital highlighted an important ethical concern associated with the genetic testing process that is the psychological impact of genetic risk disclosure. Genetic testing always carries the risk of identifying predisposing variants that are not clinically actionable, which poses a question that is continuously debated – should the result be disclosed to the patient if there is no obvious action to alleviate the condition? This group addressed this question by providing patients with background education and genetic screening for risk of coronary artery disease, and then assessed the impact on the psychology and behavioral habits of the patients. The study found that patient education resulted in a change in approach for management in a significant portion of patients that participated. With a better understanding of their monogenic and polygenic risk scores for CAD, many of these patients made informed decisions to start preventive medications based on their risk. Many of these patients also chose to undergo additional diagnostic testing. Finally, numerous patients were encouraged to take steps to better adhere to a healthy lifestyle. Within the population that were exposed to the available online educational tools, a vast majority of participants reported learning valuable information related to their health. Educating patients about their genetic risk for developing chronic disease shows great potential for impacting management of their health and encouraging patients to become proactive about preventing disease onset.

In conclusion, the 2021 ASHG Conference highlighted many new and exciting protocols and applications in clinical genomics. This article only highlights a few small nuggets out of the vast amount of valuable information presented at the conference. The constant evolution of diagnostic technology and continuous improvements in our analysis methodology repeatedly emphasize the potential impact of genomic medicine on the current state of clinical medicine. Although novel findings show great promise for future use, we continue to face many obstacles before large scale implementation of these technologies. It is essential to further our investigations in optimizing this technology to inform innovative strategies in the clinic, providing the best possible outcomes for all patients. Although NGS has great potential, its major benefits will be demonstrated when these technologies become more accessible and provide opportunities to help a larger population. Nevertheless, the future of Genomic Medicine continues to show great promise, and will soon be a gold standard in clinical care.

This article has been produced by Karan Shinde and Nicholas Wynne. Reproduction and reuse of any portion of this content requires Merogenomics Inc. permission and source acknowledgment. It is your responsibility to obtain additional permissions from the third party owners that might be cited by Merogenomics Inc. Merogenomics Inc. disclaims any responsibility for any use you make of content owned by third parties without their permission.

Products and Services Promoted by Merogenomics Inc.

Select target group for DNA testing

Healthy screening	Undiagnosed diseases	Cancer	Prenatal

Or select popular DNA test

Pharmaco-genetic gene panel	Non-invasive prenatal screening	Cancer predisposition gene panel	Full genome