Dr. Lumaka underscores the critical impact of underrepresentation in genetic studies on the ability to classify variants. “The lack of representation is a major issue. It prevents us from properly classifying certain variants, resulting in a lower diagnostic yield in our population compared to Europeans,” he explains. While many genetic variants in European populations are well-characterized, African populations face a significant data gap. This scarcity hampers the classification of novel variants and limits the ability to upgrade clinical diagnoses due to insufficient evidence.

Representation also plays a crucial role in clinical training. “If we had more faces from our community in medical literature, these patients could serve as a vital training resource for us and our students,” Dr. Lumaka notes. Representation in the literature, he emphasizes, enables clinicians to better recognize syndromes and improve diagnostic accuracy within their communities.

From a phenotyping perspective, he points out the importance of visual data. “The more African faces we have in the medical literature, the better we can identify critical phenotypes specific to our population. This representation not only aids in clinical diagnosis but also helps narrow down core phenotypes that are universally present in specific syndromes.”

Beyond equitable healthcare, increasing African genetic data enriches scientific understanding. “The more African data we have in the public domain, the better we can characterize gene functions and disease mechanisms,” Dr. Lumaka says. Integrating diverse data also helps identify shared disease phenotypes, bridging gaps between African and European presentations.

Dr. Lumaka’s research team focuses on developmental disorders, employing innovative approaches alongside traditional postnatal genetics. “We primarily see paediatric patients referred by neuro-paediatricians and decide who is likely to benefit from genetic sequencing,” he shares. However, they are now expanding into new areas, such as unique phenotyping projects and secondary findings research.

A notable breakthrough in their work involves secondary findings—unexpected genetic results unrelated to the primary diagnostic question. “We are collaborating with colleagues across Africa to deepen our understanding of secondary findings and develop context-specific guidelines,” Dr. Lumaka reveals. This initiative aims to bridge knowledge gaps and establish tailored standards for African populations.

While optimistic, Dr. Lumaka remains mindful of the challenges ahead. “Phenotyping has traditionally been the first step in selecting tests and reviewing variants,” he notes. However, as broad sequencing becomes more accessible, the reliance on phenotyping as a diagnostic tool is diminishing. “This shift introduces challenges, especially for genes without distinct phenotypic markers, making them harder for clinicians to interpret.”  He also highlights the complexities of diagnosing patients with multiple developmental disorders—a scenario more common in populations with higher rates of consanguinity. “Determining whether a patient has two distinct diseases is a critical question we must address,” he says.

Despite these hurdles, Dr. Lumaka sees a promising future driven by computational advancements and artificial intelligence. “AI and computational tools hold great potential, though the journey ahead is fraught with challenges,” he acknowledges. Overcoming these obstacles will require innovative strategies, cross-continental collaboration, and sustained investment in research and technology.

Dr. Aime Lumaka’s work exemplifies the transformative potential of genetic research tailored to underrepresented populations. By increasing African representation in genomic studies, fostering breakthroughs in phenotyping, and embracing cutting-edge technologies, he envisions a future where precision medicine becomes a reality for all. While the path ahead is filled with challenges, his dedication and vision offer hope for a more equitable and informed healthcare landscape in Africa.

Explore Dr. Aimee Lumaka’s latest research on the application of Face2Gene in African populations. This study evaluates the tool’s performance in recognition of the fragile X syndrome gestalt in Congolese subjects. Read the full study here: https://www.sciencedirect.com/science/article/abs/pii/S1769721223001258?via%3Dihub

The post Genetic Research in Africa: An Interview with Dr. Aime Lumaka appeared first on FDNA™.

At FDNA, we found ourselves amidst, and on track with, the seemingly daily advancements in the health and technology sphere. Our small (but mighty!) team worked tirelessly in 2019, closing out the year and decade one step closer to our ultimate goal: ending the diagnostic journey for those living with rare genetic diseases. With the help of our existing global network of healthcare professionals, in addition to 25,300 new users of Face2Gene, we were able to impact the lives of 158,595 new patients this year.

From forging innovative partnerships to launching cutting-edge AI-based technologies, here are some of our highlights from 2019:

• Launched Face2Gene LABS, a new and exciting product that significantly improves the diagnostic efficiency and accuracy of NGS
• Entered into a strategic partnership with PerkinElmer Genomics, the first global genetic testing lab to fully integrate next-generation phenotyping into their genetic analysis workflow
• Pioneered an innovative path for rare disease clinical trial recruitment in collaboration with Covance
• Co-authored 15 new peer-reviewed publications in high-impact scientific journals supporting our NGP technology as a gold standard in clinical genetics, including:
• Identifying facial phenotypes of genetic disorders using deep learning, Nature Medicine
• PEDIA: prioritization of exome data by image analysis, Genetics in Medicine
• Continued to advocate for the ethical use of AI in healthcare, including a featured talk at HLTH 2019
• Recognized as an opinion leader in the field of AI and facial analysis in healthcare by 500+ news publications, including the Wall Street Journal, Science Magazine, Forbes, and more

As we step into 2020, we are staying ahead of the curve when it comes to healthcare and technology trends, and how FDNA’s mission fits into this ever-evolving landscape. With AI-supported diagnosis through medical imaging, mainstream genome analysis and a call for more responsible monitoring of technology on the rise, we are primed to meet the challenges this next year, decade, and chapter of innovation has in store. 2020 is poised to take us even closer to bringing genetics to the forefront of mainstream medicine.

If you are interested in learning more about how to be part of our journey, please reach out to info@fdna.com.

The post 2019 Highlights appeared first on FDNA™.

That’s host Simon Smith’s takeaway from the inaugural episode of BenchSci’s Artificial Intelligence in Drug Discovery podcast, a show offering knowledge and inspiration from leaders in the healthcare, tech, and pharma industries.

The Role of AI in Modern Drug Discovery

This episode featured FDNA’s CEO, Dekel Gelbman, who discussed the application of current technologies (like FDNA’s Face2Gene) to identify disease-causing genetic variants and disease mechanisms when genomic data alone can fall short of a definitive diagnosis. The abundance of available data can provide better insights and transform care, but no human alone could ever effectively analyze and decipher that level of information. “Genetics really needs this type of technology as a complementary technology in genetic analysis and variant analysis,” Dekel explained.

“Using deep learning and facial analysis, clinicians have improved workflows that allow them to better treat disease,” he went on to say, “and can provide insights to possible diseases that they may have traditionally never considered.”

Through a better understanding of AI and how it can be used in a targeted way, companies can more easily collect and interpret data, ultimately leading to earlier diagnosis of disease and discovery of new ones. With this information, pharmaceutical companies have the ability to research and develop more precise treatments for these diseases and improve the quality of life for patients all over the world. From machine learning to biomedical research, digital health and pharma companies are using AI to speed discovery and cut costs, transforming drug development with data and algorithms.

“Once you understand the biological mechanisms, once you understand what is causing a particular phenotype, we’re in a much better position to try and match those diseases with specific biological approaches.”

The post Artificial Intelligence in Drug Discovery: A Conversation with Dekel Gelbman, CEO of FDNA appeared first on FDNA™.

“An early diagnosis for a child with Sanfilippo Syndrome can literally mean the difference between a chance at survival and certain death,” said Cara O’Neill, Scientific Director of Cure Sanfilippo.  “We have spoken to families who have missed enrollment criteria due to age or disease progression, or even more sadly, they met criteria, but the trial did not have remaining spots to accept them.  The number of patients included in clinical trials have historically been extremely small, sometimes taking as few as four patients.”

Those moments of powerlessness in the face of a rare diagnosis is something no parent or patient wants to feel. While in the past, advocacy groups have channeled their community’s hunger for involvement through fundraising and awareness campaigns—they are now utilizing the help of outside labs, life-science companies, researchers and online registries to turn their patient experience into viable research material.

Patient organizations like Cure Sanfilippo, Jonah’s Just Begun, and The Focus Foundation are all part of the Genomics Collaborative®, an FDNA initiative which seeks input from patients, advocates, and medical and research communities. Together, their knowledge and case examples update our understanding of different conditions and ultimately help to shorten the length of an undiagnosed patient’s diagnostic odyssey.

The Genomics Collaborative launched on World Rare Disease Day 2018. It held an open call to all corners of the community to get involved in research opportunities. Clinics and labs, universities, and patient advocacy organizations have all started using FDNA’s technology as a way to improve on the data they collect on their unique disease populations.

With opportunities to open secure, online patient portals, health organizations, and advocacy groups can enter a child’s front-facing photo to be analyzed by these technologies in an effort to develop a computational analysis of facial features that relate to their specific genetic variation.

Through mass patient participation efforts, the Genomics Collaborative works to collect and analyze patient photos and notes about their disease symptoms and progression to test for similarities in each disease population. The more patients that participate, the more accurate the model for each disease becomes.

Using next-generation phenotyping (NGP) technology and artificial intelligence, FDNA’s Face2Gene program holds a growing database of over 10,000 diseases, which is used by its global network of clinicians, labs, and researchers to advance precision medicine for hundreds of millions of patients.

With its extensive and growing reach, the Genomics Collaborative is working to familiarize all doctors with a visual of what these diseases can look like in patients of varying ages, genders, and ethnicities.

“Because 49,XXXXY is such a rare disorder, having a tool such as Face2Gene trained in the recognition of the facial phenotype of this syndrome increases awareness and accessibility of an earlier diagnosis,” said Dr. Carole Samango-Sprouse, Executive Director and Chief Science Officer at The Focus Foundation, who has been working with the Genomics Collaborative. “With earlier detection of boys with 49,XXXXY, the necessary biological treatment is likely to be available and allows these boys to reach their optimal outcome.”

To get involved and see if your disease population is currently being studied through the Genomics Collaborative please visit www.genomicscollaborative.com

The post Patient Participation Proves Invaluable for Both the Diagnosed and Undiagnosed appeared first on FDNA™.

Abstract

Significant improvements in automated image analysis have been achieved over recent years, and tools are now increasingly being used in computer-assisted syndromology. However, the recognizability of the facial gestalt might depend on the syndrome and may also be confounded by the severity of the phenotype, size of available training sets, ethnicity, age, and sex. Therefore, benchmarking and comparing the performance of deep-learned classification processes is inherently difficult. For a systematic analysis of these influencing factors, we chose the lysosomal storage diseases Mucolipidosis as well as Mucopolysaccharidosis type I and II, which are known for their wide and overlapping phenotypic spectra. For a dysmorphic comparison, we used Smith-Lemli-Opitz syndrome as a metabolic disease and Nicolaides-Baraitser syndrome as another disorder that is also characterized by coarse facies. A classifier that was trained on these five cohorts, comprising 288 patients in total, achieved a mean accuracy of 62%. The performance of automated image analysis is not only significantly higher than randomly expected but also better than in previous approaches. In part, this might be explained by our large training sets. We therefore set up a simulation pipeline that is suited to analyze the effect of different potential confounders, such as cohort size, age, sex, or ethnic background on the recognizability of phenotypes. We found that the true positive rate increases for all analyzed disorders for growing cohorts (n=[10…40]) while ethnicity and sex have no significant influence. The dynamics of the accuracies strongly suggest that the maximum recognizability is a phenotype-specific value, that has not been reached yet for any of the studied disorders. This should also be a motivation to further intensify data-sharing efforts, as computer-assisted syndrome classification can still be improved by enlarging the available training sets.

The post Advances in computer-assisted syndrome recognition and differentiation in a set of metabolic disorders appeared first on FDNA™.

The post Y. Gurovich, et al. Next-Generation Phenotyping: A Performance Analysis appeared first on FDNA™.