“Our clinic is made up of a combination of five geneticists and five genetic counselors. We’re very fortunate to be able to see people of all ages throughout their lifetimes and have no real limitations on the indications that we see. We see genetics cases across the board, whether it is kids with birth problems, developmental delays or older people with neurological findings,” said GC Samantha Augustyn. “Sometimes we have a straightforward diagnosis and can do very specific testing just to confirm what we think. Other times we must do whole exome sequencing or genome sequencing because we don’t quite know what the answer is.”

“Face2Gene complements our workflow and enables us to work as a team. Obviously, not everyone is seeing everything, and not everyone is an expert in every single area, but we all have a lot of experience between us. We use Face2Gene as a collaborative tool to talk about the patient and make sure there’s nothing that we are majorly missing.”

“We take a picture of as many of our patients as possible, specifically the undiagnosed ones, and follow it through. Genetic testing is done and when the report comes back, we go back to Face2Gene to see if the relevant syndrome appeared there.”

“Knowing that Face2Gene is a protected setting that our team can have access to, we use this tool to share information between us to not only see what the algorithm gives us back in terms of possible syndromes but also all the different professionals within the team can view the clinical notes. We can share this information and ask, ‘Has anyone ever seen somebody with these features before? Here are pertinent positives, here are pertinent negatives.‘ This helps build the story and solve the problem as a team.”

This allows a conversation that can add pertinent points to what the app is showing us, almost like a new person looking at it with an outsider’s perspective, that can bring something that is important to think about.

Find out more about how to use Face2Gene as a team

Contact support@fdna.com for a demo!

The post Samantha Augustyn talks about how she uses Face2Gene as a team appeared first on FDNA™.

The post Webinar: Matching ultra-rare cases in the clinic appeared first on FDNA™.

Presented by Omar Abdul-Rahman, MD, University of Nebraska Medical Center

Precision Medicine World Conference (PMWC)| September 24-25, 2018 | Duke University

This post is based on a presentation given at PMWC Duke. Watch the full presentation.

At the Precision Medicine World Conference (PMWC) Duke meeting, Dr. Omar Abdul-Rahman, Friedland Professor and Director of Genetic Medicine at the University of Nebraska Medical Center, told the story of the emergence of phenotypic data as crucial in clinical evaluations.

According to Dr. Abdul-Rahman, there are, “three legs to a stool that we have to understand” when making a patient evaluation: genes, environment, and phenotype. He described a time about ten years ago when “there had been a lot of advancements made in the ability to get good genomic data.” However, he went on to say that those same advancements had unfortunately not been made in phenotyping.

Standardizing the Phenotype

The first step taken in tackling this standstill and beginning to capture robust phenotypic information was the development of the Elements of Morphology as a way to “standardize the nomenclature” of phenotypes. Together, a group of geneticists defined over 400 features of the face, hands, and feet to streamline they way clinicians referred to various morphologies.

“Once we standardized the phenotype, the next question became, ‘How do we capture it?’”

Capturing the Phenotype

While at a study site for the National Children’s Study, Dr. Abdul-Rahman and his team realized that they were successfully capturing genetic and environmental data, but not phenotypic. With the goal of capturing as many of the 400+ features as possible, they showed 15 photos (eight of the head/neck, four of the hands, and three of the feet) and three videos to a panel of geneticists for review.

This laborious study was presented as a poster at a conference where it was strategically placed next to a poster focused on computer-aided facial recognition. Combining Dr. Abdul-Rahman understanding of how to capture enough imaging to analyze a phenotype with his colleagues’ computer-aided automation of the process, a collaborative study between the brains behind the neighboring posters was born.

Analyzing the Phenotype

The automating software introduced, Face2Gene is a suit of phenotyping applications developed by FDNA that operates on AI and deep learning technologies. Using facial analysis, the system has evaluated over 150,000 patients, and cross-comparing their phenotypes to a growing database of over 10,000 genetic diseases helps to enable more rapid and accurate diagnoses.

Studying FASD with Face2Gene

After learning about Face2Gene, Dr. Rahman chose to apply the technology to a particular teratogen of interest, alcohol, and related Fetal Alcohol Spectrum Disorders (FASD). FASD can be broken up into four diagnostic categories, the first three of which are straightforward:

• FAS – requires all of the typically-present features
• Partial FAS – requires some, but not all of the typically-present features
• Alcohol-related birth defects (ARBD) – includes a series of anomalies more common to present among alcohol consumers
• Alcohol-related neurodevelopmental disorder (ARND)

The final FASD category, Alcohol-related neurodevelopmental disorder (ARND), is more difficult to diagnose because there are no outward features. When diagnosing, there are two requirements: documented prenatal alcohol consumption and neurobehavior impairment, which cannot be diagnosed under the age of three.

Because ARND is both the most difficult FASD to diagnose and also the most prevalent FASD type in the United States, Dr. Abdul-Rahman decided to put Face2Gene to the test to see if the technology could detect the difference between the four diagnostic categories. He ran images of over 130 subjects with FAS, Partial FAS, ARBD, or ARND, as well as controls through Face2Gene in a cross-validation test. The images were split 50/50 into training and test sets and were run through ten rounds. The results showed that for any FASD vs. control, the manual and computer-aided score both performed well and relatively equally. The same was true of looking at the individual syndrome categories, FAS, Partial FAS, and ARBD; however, when looking at ARND vs. control, there was a better AUC for computer-aided vs. manual.

You can find the study here. 

Lessons Learned

Although there appeared to be no physical characteristics in the clinical criteria for ARND, Face2Gene was able to pick up on subtle facial cues, showing that, “there must be something that is broader, but is really subclinical from a provider standpoint.” Dr. Abdul-Rahman concluded with the following lessons learned from the study and his experience using Face2Gene:

• Facial analysis software like FDNA’s Face2Gene can be used to delineate phenotypes at a subclinical level
• Such software may allow for the identification of external physical biomarkers for conditions such as FASD
• Such software can potentially be used by primary care providers as a screen for further investigations (e.g. genetics referral)

Check out other presentations and similar research on the benefits of AI and facial analysis in personalized medicine.

The post Personalizing Medicine with Artificial Intelligence and Facial Analysis appeared first on FDNA™.

Cara O’Neill, MD, FAAP

VP, Scientific Director, Cure Sanfilippo Foundation

Cara O’Neill’s daughter, Eliza, was diagnosed with Sanfilippo about five years ago, which thrust her abruptly into the advocacy sphere.

Sanfilippo, also known as mucopolysaccharidosis III (MPS III), is a lysosomal storage disease caused by an enzyme deficiency that leads to a buildup of heparan sulfate, a glycosaminoglycan. The buildup impacts the central nervous system and causes progressive neurodegeneration; Sanfilippo has been referred to as childhood Alzheimer’s because of the symptoms it causes.

O’Neill, a pediatrician herself, quickly recognized the issue of diagnostic delay in identifying the syndrome in patients.

“Newborn screening wasn’t just right around the corner,” she said, which is why she was excited to meet FDNA at a genetics conference. “We thought, wow. What a great tool to help us find kids earlier.”

Cure Sanfilippo Foundation (co-founded by O’Neill and her husband) and FDNA set to work training the Face2Gene system on Sanfilippo syndrome. After creating a patient portal for parents to submit photos, the artificial intelligence powering Face2Gene eventually learned to differentiate the faces of patients with MPS III from faces of unaffected controls, as well as distinguish between MPS III and other MPS syndromes.

O’Neill also wanted to see how patient faces changed over time, so she gathered cohorts based on age and created average faces for each.

“What was really exciting to me about this was that we were able to pull out that early age group in the one to three-year-old range where symptoms are starting to become more apparent to doctors and parents. We thought this is where we could possibly pick up children early and the software was able to distinctly, with 95% accuracy, recognize MPS IIIB patients in the 1-3-year-old age range from other syndromes and from unaffected controls. That really validated what we had hoped this tool might be for us,” she said.

O’Neill said her experience as a parent and pediatrician impacted her desire to make other pediatricians more comfortable with disorders like MPS III.

“We would like to identify these kids before they have so many symptoms that they make it to the specialists.”

As part of that effort, Cure Sanfilippo and Greenwood Genetic Center partnered to create a set of trigger parameters prompting pediatricians to use Face2Gene to seek diagnostic consultation with genetic experts.

The process has been extremely rewarding for O’Neill, who had not done this kind of research work before. “It was surprisingly easy to move these projects forward,” she said, encouraging others to give it a try.

Karen Gripp, MD, FAAP, FACMG

Chief, Division of Medical Genetics, A.I. DuPont Hospital for Children

Chief Medical Officer, FDNA

Dr. Karen Gripp uses Face2Gene in both clinical and research settings. With one particular patient, she described how with just a single photo and seven features, Face2Gene presented Smith-Lemli-Opitz as a potential match, which allowed Gripp to order targeted testing rather than whole exome sequencing.

Gripp also used Face2Gene to identify a “face” for Ayme-Gripp, a syndrome previously described as “Down-syndrome-like.” “There was a long ongoing debate about how distinct a condition is it,” Gripp said.

Gripp noted how a facial composite of Ayme-Gripp syndrome is distinct: small eyes, short nasal tips, and full eyebrows are clearly evident. She was able to use Face2Gene to prove that these “average faces” are indeed distinct, different from those of children with Down syndrome.The system analyzed the facial phenotypes of the three separate cohorts with binary comparisons and multiclass comparisons. “For each photograph, the system has to decide which bucket to place it in,” she explained.

Gripp said she appreciated the RESEARCH application of Face2Gene for the way it streamlines hypothesis testing. “You get statistical data that is very helpful in terms of proving your impression in an objective manner.”

Ilana Jacqueline

Patient Advocacy Manager & Genomics Collaborative Project Coordinator, FDNA

Ilana Jacqueline said after the 2017 Year of Discovery that she and her colleagues realized “a year of discovery is not enough” and used the momentum from their monthly deep dives into specific diseases to build the Genomic Collaborative.

“Our technology continues to learn by way of collaboration,” Jacqueline said, explaining how in Face2Gene, “facial features become a mathematical descriptor,” which maintains patient privacy while optimising system learning. “Ultimately that reduces the amount of testing that may need to be done to confirm a diagnosis,” she said.

Clinicians and researchers who join the Genomics Collaborative can access forums to discuss difficult cases and gain the support of the FDNA research team to support the creation and publications of scientific papers, while of course maintaining ownership of their discoveries. As a bonus, because the technology is free, doctors and advocacy groups can devote their resources to other projects instead.

As a result, patients and families can see growing communities from improved diagnostic rates. “Patients will gain a greater understanding of their diseases as more experts become involved in the research,” she said.

“The need is urgent always,” said Jacqueline, who understands this from personal experience as a rare disease patient. “You’re dealing with families and children whose survival is a ticking clock. You know the stakes are high.” That’s why, she said, “We should always be learning.”

The post The Genomics Collaborative: How to Design the Future of Health—Together appeared first on FDNA™.

By now, AI is so commonplace it’s almost embarrassing if a tech company doesn’t use it. But as much as the sci-fi world likes to pretend otherwise, truly strong general AI is still far away; most AI solutions involve machines programmed for very specific tasks. The most complex systems may conduct many of those tasks, but they’re still not cognitively independent. They still need humans for the most complex problem-solving, and nowhere is that more evident than in medicine.

While computers can far outpace humans as processing and analyzing big data, the human touch is still essential to assessing the whole and achieving larger goals. Educating consumers about the line between man and machine is a crucial part of corporate responsibility in the AI space.

Take for example Face2Gene, FDNA’s flagship suite of phenotyping applications. At best, it suggests ideas that never would have occurred to a clinician. At worst, it can become a crutch. FDNA uses regular “Name This Syndrome” challenges to gamify the use of our tools and to remind users to question AI. Follow your GPS exactly and you might drive straight into a ditch–the same is true for relying too heavily on machine learning in the clinical diagnosis process.

In addition to educating users and consumers about the limits of AI, companies must consider how they can influence society. Will AI be only a tool for the wealthy, the white, and the west? Or can AI level the playing field?

The “garbage in, garbage out” premise applies to inequality as well. “Disparities in, disparities out,” we could say of AI data inputs and resulting analyses. (It’s hard to forget the stories of AI bots or algorithms becoming racist or disturbingly morbid after learning from the masses on Twitter and Reddit.)

In the genomics world, the majority of data comes from patients of European descent. To combat that uneven distribution of data, FDNA has worked diligently to create a global network of users and sites in over 130 countries. We have both a free tool and software as a service. Four years after our product launch, more than 50 percent of the patient data contributed is non-Caucasian. Any AI platform is only as good as what it’s being trained on, so as you might expect, when we feed the system with more diverse data, it becomes more robust.

Of course, the more data a company or tool touches, the more concern users have regarding privacy. Fortunately for FDNA, a byproduct of computer vision is de-identified photos, so we’re able to share the information from the data without actually sharing the data. This allows us to spread the benefits of our insights while still protecting patient privacy.

We as companies in this space can learn from one another how to best protect and serve our users and customers. Precision medicine may be powered by AI, but the best applications of these tools require man and machine.

The post AI in Healthcare: Separating the Hype from the Hope appeared first on FDNA™.

This post is based on a presentation given at the Precision Medicine World Conference. You can skip to the video by clicking here.

As researchers and clinicians expand their investigation of the relationship between phenotypes and genotypes, the importance of accurate phenotypic analysis grows. That means the field needs a uniform understanding of what phenotypes are, how to describe them, and how to assign traits.

In differentiating deep phenotyping and next-generation phenotyping, Dr. Donald Basel (Medical Director of the Genetics Center at the Children’s Hospital of Wisconsin) asked Precision Medicine World Conference attendees, “Where do we limit our thoughts as to what a phenotype represents? Is it limited to structural morphology or do we consider all aspects of the phenotype?”

Deep phenotyping, which has existed since the 1950s, made a leap forward in dysmorphology when in 2009 NHGRI developed a common language, the hierarchical Human Phenotype Ontology (HPO). Next-generation phenotyping expands that concept, including not just structural morphology but aspects like speech analysis, gait analysis, and all of the “omics”–metabolomics, microbiomics, etc.

Even with the benefits of a common language, clinicians are still faced with the challenge of properly applying the HPO terms. Dr. Basel (also a member of the FDNA Scientific Advisory Board) noted that a tool like Face2Gene, a suite of phenotyping applications, does some of this identification for the clinician, and gave an example of human vs machine analysis of Cornelia de Lange syndrome and phenotypically similar syndromes. A panel of experts identified patients with Cornelia de Lange 77 percent of the time and spotted similar syndromes 87 percent of the time, with a clinical sensitivity of 82 percent and a specificity of 89 percent, whereas FDNA’s DeepGestalt technology correctly selected Cornelia de Lange 94 percent of the time and similar phenotypes 100 percent of the time, with a sensitivity of 86 percent and a specificity of 100 percent.

Of course, as Dr. Basel pointed out when it comes to Cornelia de Lange, a well-trained dysmorphologist can “essentially walk into a baby’s room and make this diagnosis.” The more profound results are for patients with phenotypes that would be far more difficult to diagnose because of the subtlety of facial traits, or for helping non-dysmorphologists recognize these phenotypes.

In a March 2018 publication, Dr. Basel described as “quite daring,” researchers took four syndromes with inborn metabolic errors and mild facial feature coarsening, plus Nicolaides-Baraitser syndrome (which has a similar phenotype) to test the

“Just looking at pure facial imaging the software was able to accurately predict the specific disorder in 64 percent of the cases. If you added a single feature into that feature set, the diagnostic accuracy increased to 87 percent.”

This means clinicians using Face2Gene may be presented with syndrome recommendations that they might have otherwise disregarded.

Similarly to how clinicians gain expertise with practice, Face2Gene grows smarter as it sees more cases. “The more we use it, the better it gets,” Basel said.

This brought him back to his original question: if artificial intelligence technologies can grow to analyze more and more data of increasingly varied types, “Where are the limits?”

Donald Basel presents at PMWC

The post “Where are the limits?” appeared first on FDNA™.

A flock of researchers from around the globe shared their findings in dysmorphology and molecular genetics at this year’s ESHG as a part of FDNA’s corporate satellite talk and various scientific posters.

Karin Weiss (Rambam Health Care Campus, Haifa, Israel) presented her further work on Sifrim Hitz Weiss Syndrome (SIHIWES), a recently described form of syndromic intellectual disability identified through reverse phenotyping.

“When you come to classify the new case, there are some difficulties because the phenotype is not specific enough to make the diagnosis,” she said, adding “in this condition all the variants are missense.”

Weiss attempted to train Face2Gene on known SIHIWES cases and successfully taught the system to separate healthy individuals from those with SIHIWES, although the rate at which the system separates SIHIWES from different syndromes was not statistically significant. However, the system was able to notice the possibility of SIHIWES in two patients whose exome sequencing indicated de novo mutations outside the “hotspot” on the CHD4 gene, leading her to conclude that, “facial recognition can aid in variant interpretation probably by supporting a specific variant, not excluding.”

Antonio Martinez-Monseny, MD, (Hospital Sant Joan de Deu, Barcelona, Spain) reviewed how his team used Face2Gene to test several hypotheses, first training the system on PMM2-CDG with a mean accuracy of about 75 percent when comparing confirmed cases to unaffected controls and diagnosed cases of Angelman syndrome. He confirmed that there is indeed a “face” for the syndrome and that it is recognizable across ages, making it easier for clinicians to diagnose the syndrome at an earlier age and, he said, potentially improving access to therapies.

Peter Krawitz, MD, PhD (University of Bonn, Germany), explained how a web-based filtering and annotation tool, Deep Phenotyping, Deep Learning (DPDL) is powering his Prioritization of Exome Data by Image Analysis (PEDIA) approach with help from Face2Gene.

“DPDL can integrate multiple scores from the molecular and phenotypic level,” he noted. Currently, clinicians can use DPDL via the website, but it will soon be available through Face2Gene LABS as well.

Krawitz, who is also FDNA’s Chief Data Science Officer, emphasized how combining phenotypic and molecular information refines the results either could give alone.

“In a routine setting this will speed up your analysis from maybe several days to several hours,” he said.

Jean Tori Pantel (Charité – Universitätsmedizin Berlin, Germany) demonstrated the RESEARCH application with Face2Gene, which she is using as part of her thesis investigating computer vision applications for recognizing inborn errors of metabolism. Her lab was able to create separate masks (composite facial photos) for types of mucopolysaccharidoses that were previously undifferentiated. In the process, she created cohorts of varying size with and without confounding factors, eventually concluding that the distinguishability of cohorts improves as the cohort sizes increase, although the achievable maximum is still unclear.

Tzung Hsieh (Institute for Genomic Statistics and Bioinformatics, Bonn, Germany), whose poster was nominated for a “Best Poster” award, researched how patients with mutations in similar molecular pathways compare phenotypically.

Idan Menashe, PhD (Ben Gurion University of the Negev, Beersheva, Israel), shared his research on facial dysmorphisms as biomarkers for autism spectrum disorder. Using a cohort of 81 patients at the Negev Autism Center, Menashe and his team used the Face2Gene deep convolutional neural network to evaluate the photos, as well as those of controls, and compare the average faces of the case group and the control group. The groups showed clear separation, with a p-value less than 0.001. His team also evaluated the specific areas responsible for the separation and concluded that the upper facial area (eyes and nose) are most informative to the system.

In addition to these presentations,  Yaron Gurovich,  Karen Gripp, MD,  and Ben Pode-Shakked, MD had scientific posters at ESHG showcasing how Face2Gene can be applied to research and how FDNA is improving the suite of phenotyping applications.

The post FDNA and Face2Gene Featured at ESHG appeared first on FDNA™.

Dr. Christine Stanley, Head of Clinical Laboratory for WuXi NextCODE, US and CLIA medical director for Q&A Diagnostics, discusses next-generation phenotyping for improved variant interpretation through integration of Face2Gene LABS with WuXi NextCODE’s variant interpretation system.

According to Dr. Stanley, the ACMG has set standards on the classification and interpretation of genetic variants, including various levels of classifications as benign, unknown or pathogenic.

There are 28 lines of evidence used for classifying these variants.

“When you look at the phenotypic line of evidence, it wasn’t one that was utilized often […] but it really could be if we make it better and more complete,” she said. “In the ACMG publication, to use the phenotype as a line of evidence in variant scoring, you really need to have high clinical sensitivity, the patient must have a well-defined clinical presentation, the gene is not subject to a lot of variation, and the family history is consistent with the mode of inheritance of the disorder.”

“The phenotype information is really critical to performing whole exome and genome interpretation,” she continued.

Dr. Stanley goes on to explain that the phenotypic information allows Wuxi to utilize this phenotypic line of evidence in their variant classification and to help reclassify variants of unknown significance into the pathogenic category.

However, according to Dr. Stanley, phenotypic information is often lacking or not provided to support NGS interpretation.

She goes on to describe how WGS or WES produces a filtered list of variants that have clinical relevance, and that the report is static from what was known at the time. According to Dr. Stanley, “what we need are tools that allow for revisable reporting” because new information—such as new understanding of gene variations, or new patient symptoms—can become known down the road. This new information, ideally, can be used to reanalyze past genomics test results, impacting medical management, clinical research or patient support.

“We really need digital tools to get there,” she said. “We have taken the FDNA tool [Face2Gene] that captures the detailed clinical phenotype, and we’ve incorporated it into WuXi’s clinical sequence analyzer in order to do that real-time variant review.”

Dr. Stanley goes on to describe how the phenotype can be broadened to include other data, such as medical imaging, biometrics, clinical notes, and more, all pulled into Face2Gene and integrated with the variant interpretation.

She demonstrates how clinical users capture photos, anthropometric measurements, phenotypic features, and more within Face2Gene, and the application utilizes artificial intelligence to assist the user in further annotating the phenotype, resulting in a comprehensive phenotype for the patient. From there, the clinician can pass the phenotypic data securely to WuXi by selecting WuXi from the list of labs, resulting in the data being transferred into Wuxi’s clinical sequence analyzer.

Wuxi then reviews the gene list associated with the patient’s genetic sequence, scores the variants based on the ACMG classification criteria, and uses the clinical phenotype to match up to and highlight the variant results, or to highlight variants of unknown significance that relate to the phenotype to help determine if reclassification makes sense.

Watch the recording of the talk On Our Channel.

The post Dr. Christine Stanley: Face2Gene LABS at WuXi NextCODE: Phenotyping for Improved Variant Prioritization appeared first on FDNA™.

HISTORY

Since the 1960s, the field of medical genetics has been evolving rapidly with regard to phenotype delineation and analysis.

A disease phenotype requires a multi-faceted analysis. Besides defining the phenotype with diagnostic criteria, medical geneticists are also faced with the challenge of drawing lines along spectra of manifestations and within disease definitions, allowing for how symptoms may change over time.

To measure these variables, biochemical analysis, detailed histories, anthropometrics, standardized measures, and medical imaging can be applied in concert. As next-generation sequencing evolves the way the medical field examines patients’ genomes, next-generation phenotyping integrates these changes into the analysis of human health, maximizing the impact of new sequencing technologies.

“We’ve actually entered a new golden era of phenotyping,” Dr. Carey said.

PHENOTYPIC DOMAINS

Dr. Carey noted there are three domains to measure disease: diagnostic criteria (a.k.a. definition of the phenotype), the spectrum of manifestations and complications, and the natural history of how the phenotype changes over time.

SYNDROME DELINEATION

Syndrome delineation has three distinct parts, as Dr. Carey reviewed in Charlotte, N.C. The first, the physical examination, involves observations and descriptions of the patients. Then there are two levels of syndrome genesis to consider: formal (e.g., a shortage or malformation of a critical enzyme) and causal (e.g., the genetic source of the formal genesis, such as a deletion or de novo mutation.)

SYNDROME GROUPS & HETEROGENEITY

The broad definition of syndrome classification is misleadingly simple: simply group patterns of anomalies, with at least one being morphologic, thought to be etiologically related. Of course, in practice, this is very different than in theory; groups of syndromes, variant forms, subtypes, or related disorders can be hard to separate into discrete syndromes. Similar phenotypes can be caused by errors in different genes, or by multiple genes, and identical mutations can cause different phenotypes across patients.

“There are syndromes we know quite well that have 17 genes. There’s one gene that has 17 syndromes. Do we go with the gene or do we go with the phenotype?” Dr. Carey asked the audience.

THE AXIS MODEL

To solve these discrepancies, Dr. Carey proposed the use of the “axis model.” Axis I describes the clinical phenotype, Axis II describes the underlying molecular genetics, and Axis III describes non-genetic factors, like the environment.

“It’s never been adapted but I actually like it,” he said.

I would propose that even though you can’t [describe a syndrome in] 3 or 4 words, or less, Dr. Carey said of the sometimes wordy model. Despite its potential complication, its focus on phenotype keeps the naming mechanism patient-centered.

Watch the recording of the talk On our Channel.

The post Dr. John Carey: Delineating Genetic Syndromes and Next-Generation Phenotyping appeared first on FDNA™.

Starting in 1996, a select few clinicians began identifying a small group of patients with similar phenotypic features: Down-syndrome-like facial features, short stature, intellectual disability, cataracts and sensorineural hearing loss. Among these clinicians was Dr. Karen Gripp, who over the next two decades would become one of the namesakes for this MAF transcription factor related rare disease: Aymé-Gripp syndrome (AGS).

In her research, Dr. Gripp (the new Chief Medical Officer for FDNA) asked, “Does Aymé-Gripp syndrome have a facial phenotype that is recognizable using automated facial analysis?” Last week, she shared her investigation at the ACMG annual meeting.

Dr. Gripp used the Face2Gene RESEARCH application to create composite facial images for AGS and for Down syndrome, as well as for individuals with no suspected syndrome.

She ran the analysis twice, using new sets of cases for the Down syndrome and control cohorts for the second analysis to account for the “possibly subtle difference” in averaged facial images.

“There are twenty different images in those cohorts,” Dr. Gripp explained.

Dr. Gripp described the results provided by Face2Gene, including a multiclass comparison that reveals how well the system can successfully classify patients into the correct diagnosis based on facial analysis.

For each analysis, Face2Gene RESEARCH also provides statistics on the performance of the system.

“What you look at here is the mean accuracy for the analysis, and you have to always put that in perspective to the random chance for that particular analysis,” said Dr. Gripp.

Dr. Gripp’s use of the RESEARCH application is easily repeatable by any scientist on their “favorite syndrome.”

First, researchers should upload, to Face2Gene CLINIC, any patients (with at least one frontal facial photo) that they plan to use for research cohorts, being careful to clearly label [in the “case name” field] the study cohort that the patient should be added to, for example: “John Doe, Down Syndrome Cohort.” This makes it easier to find the relevant cases later when creating your study cohorts. If you are interested in analyzing the phenotype distribution for the disease, it is also important to thoroughly annotate a defined set of those phenotypes (present or absent) for every case. Otherwise, just a frontal facial photo for each case is sufficient for the analysis.

Once all the cases are added to Face2Gene CLINIC, you can move to the Face2Gene RESEARCH application by clicking “RESEARCH” in the top right corner of the app. There, you can create a new project, adding the cohorts you want to compare. To do this, select “add cohort,” and select the relevant cases from the list. Repeat for each cohort you want to create, making sure each cohort is roughly the same size.

“If you have a huge imbalance in the cohort size, that by itself introduces a bias,” Dr. Gripp mentioned.

Once cases are uploaded and cohorts are created, users can run their experiment with a click.

“You push the button, you get yourself a cup of coffee; it takes a little bit, literally a few minutes, for the system to run the analysis.”

When the analysis is complete, the user receives an email with summary results and can look in more detail at each item. Results include the sensitivity and specificity reported as the area under the curve of the Receiver Operating Characteristic (ROC) curve, and the p-value to measure significance.

All of the charts and visualizations are easily copied for use in papers and publications by the researcher.

For Dr. Gripp, using the Face2Gene application resulted in the addition of a new syndrome within Face2Gene, paving the way for clinicians globally to better recognize the AGS phenotype in patients through use of  Face2Gene–and any researcher with an internet connection can now do the same for other diseases.

Watch the recording of the talk On our Youtube Channel.

The post Dr. Karen Gripp: Face2Gene RESEARCH for Deep Phenotyping of Novel Syndromes appeared first on FDNA™.