A Needle in the Haystack: On Inverted Nipples and Looking in All the Wrong Places

Oct 8

On October 7, 2026, after 16 years of looking, I learned that my son Charlie has a variant in SLC35A2, a gene associated with SLC35A2-congenital disorder of glycosylation (SLC35A2-CDG). Charlie developed infantile spasms before his first birthday. Since then he has had ongoing seizures, severe intellectual disability, two brain surgeries, and many rounds of genetic testing that found nothing.

My work is in data science and biomedical informatics. Part of that work is figuring out how to document, in meticulous detail, people with mysterious diseases such as amyotrophic lateral sclerosis (ALS), where the cause is often unknown, so I know as well as anyone how complicated and challenging it is. I also have the lived experience of searching for my own child's diagnosis. These are my first impressions, one day after getting the answer, of what I learned.

The features that were never written down

The first thing I learned is that the features with no obvious connection to the disease were as important as the seizures, and in Charlie's case probably more important. In epilepsy we tend to focus on seizure types, seizure frequency, age of onset, developmental concerns, then run those data alongside the genetic results and look for a match. Over 16 years of studies, those were the things I was asked about, and they are important.

Charlie is a handsome, beautiful child, and I would not describe him as having facial differences. He does have thick eyebrows, a broad nasal bridge, eyes with a slightly unusual shape, and inverted nipples. I began telling anyone who would listen about the nipples before he turned one, soon after the spasms started, and the usual reply was to ask why I was worried about his nipples. Experts shrugged, called it a quirk, and did not write it down, and I now believe that is where the search kept going wrong.

After the result came in, I read the SLC35A2-CDG page on the National Organization for Rare Disorders (NORD) website. Its overview lists thick eyebrows, a broad nasal bridge, and inverted nipples among the dysmorphic features, next to early epileptic encephalopathy, severe developmental delay, and brain abnormalities on magnetic resonance imaging (MRI). Those are the features I had been pointing to since Charlie was a baby. Across two brain surgeries and countless studies, nobody had asked me about them (although I would make sure they were noted down anyway), and now they read as an obvious clue.

Those features help because of how a variant of uncertain significance is resolved. The team compares the child's phenotype with published cases of other people who have variants in the same gene, and asks how likely the resemblance would be by chance. Epilepsy and intellectual disability appear in a great many genetic conditions, so on their own they do little to separate one gene from another. If inverted nipples are also reported in children with variants in that gene, then a child with epilepsy, intellectual disability, and inverted nipples looks much more like a match, and the team can be more confident in the diagnosis.

That comparison is only possible when somebody has recorded the feature. For us, it took a clinician sitting down with Charlie and writing down everything, in full detail, before the pieces could be compared.

How the answer was found in brain tissue

The variant was found in brain tissue removed during one of Charlie's epilepsy surgeries, which is a longer story that I will share separately. For years I understood that the tissue could not be tested, because of the way the sample had been fixed and stored after surgery and the methods available the last time we looked into it.

At a conference, I sat down to lunch and happened to start talking with a doctor I had never met. I was ready for that conversation because I give everyone the same spiel about Charlie, and I had it prepared. When I mentioned how unfortunate it was that the tissue could not be tested, he told me he thought there might be a workaround. We talked it through, I found someone to run the testing, and I arranged for the samples to be sent. Having my story ready is what allowed me to meet the moment and seize the opportunity, and without that conversation we would still not know Charlie's cause.

What I would tell another family

Trust your gut. My gut told me from the beginning that when we figured out what was wrong with Charlie, it would have something to do with his inverted nipples. Plenty of people told me I was focusing on the wrong things, and most of those who found it interesting did not take it seriously. I was not a panicking, grieving mother who needed to be humored. I could see that he had a constellation of features, and I had never seen another child who looked like him. To be clear, after speaking to the genetic counselor, it seems that these little quirks were not the main basis for the diagnosis - they relied heavily on more traditional neurology-related features. However, it is clear that if inverted nipples, bushy eyebrows, and the rest are so common, that flagging these features in kids who are still undiagnosed would be a great help in the analysis.

Find a clinical champion. We have an incredible neurologist who did what it took. Becoming a self-taught genetic counselor was not on anyone's bingo card when their child was born, so we may not describe what we see in the expected vocabulary. A champion is someone who is willing to work out what you are trying to say, and who appreciates your persistence about the things your gut tells you.

Look for team science. Charlie's neurologist was open-minded about working with a different neurologist on this, and I think it was an incredible model of team clinical care and of team science. We would not have an answer without it, because the doctors could have disagreed on the approach and set my wishes aside, and then we would never have known.

Talk to the younger clinicians and scientists. They know the newer methods and they have good ideas. Our answer came from a young doctor with a new idea, and I would have missed it if I had only listened to the people with the most seniority.

Be the squeaky wheel. Call everyone and talk to everyone, without being intimidated and without worrying about sounding stupid. We would not have an answer if I had not struck up a conversation with a stranger over lunch.

Have your story ready, unusual features included. You may meet the right person with no warning, so prepare a short version you can say on the spot. Mine was that my son had late-onset infantile spasms, that he has a diffuse brain malformation unlike the classic patterns clinicians are used to seeing, and that he has a broad nose, bushy eyebrows, a slightly unusual facial appearance, and inverted nipples.

Learn the basics of genetics, and help your community learn them. In the 24 hours after I got Charlie's result, I spoke with parents, people who work in this field, and people who run patient advocacy groups who did not know that a variant can be somatic instead of germline, or that mosaicism means a variant can be present in some of a person's cells and tissues and absent from others. Our answer was in brain tissue after years of other testing, which is why those concepts are on my mind. It is hard to advocate for something you do not understand, so I see a real opportunity to educate ourselves and our communities about genetics, and about how detailed phenotyping is used to interpret a genetic result.

Open science and data standards for our disease communities

Our disease communities need to embrace open science, data standards, and ontologies. The people behind efforts such as the Mondo Disease Ontology and the Global Alliance for Genomics and Health (GA4GH) know how to document a phenotype so that it can be compared with other cases, how to share data, and how to write the software that does the comparing. Many families and advocacy groups have never heard of informatics, so this work gets little of their attention, yet these are people who want to help us and have the skills to do it. What they taught me is what allowed me to push Charlie's case forward and be ready when the opportunity came.

Our own answer shows the connection: the NORD page where I recognized Charlie states that its content was gathered in partnership with the Mondo Disease Ontology. The Human Phenotype Ontology (HPO) has a term for inverted nipples, HP:0003186, so a clinician who notices the feature has a standard, computable way to record it.

I think one of the best things a community can do now is get its data ready for artificial intelligence (AI), so that classifying each person correctly gets easier and mistakes get harder to make. The standards I rely on for that are HPO, Mondo, GA4GH Phenopackets, and the Observational Medical Outcomes Partnership (OMOP) Common Data Model, and I am now learning the Linked Data Modeling Language (LinkML). Links to each are at the end of this post.

Registries: HPO-based questions and links to biological samples

If you are building a registry, I recommend including, at a minimum, a general health survey whose questions are attached to HPO terms. Many large registries use the GenomeConnect survey or a similar HPO-based survey. GenomeConnect is the patient registry of the Clinical Genome Resource (ClinGen), and Savatt and colleagues describe it in a 2018 paper in Human Mutation, which reports that other registries and research groups have adopted all or parts of the survey. The supplement to that paper shows the question wording, the HPO mapping for each question, and the branching logic.

The survey works as a review of systems. It asks about each general body system, and a participant sees the follow-up questions for a system only after reporting an issue there. That design lowers participant burden while still covering the broad areas where a person might be affected, and I understand that several popular registry platforms, including Across Healthcare's Matrix platform uses the same approach.

This coverage is especially important in studies where researchers receive the data downstream and cannot go back to participants with questions. Your doctor is not behind the scenes with the researchers who use the data, so they have only what was reported. The more standardized, structured, and comprehensive that report is, the better the chance that someone can find the pattern, because a review of systems asked of every participant gives a feature outside the main disease area a place to be recorded.

Patient advocacy groups that create their own registries also need to know where their participants' biological samples and biological data are held. Our answer came from tissue stored after a surgery years earlier, and it was only useful because someone could find it and test it. I would encourage groups to collect that information and to make sure the phenotyping from a natural history study can be paired with the biological data, since each is much harder to interpret without the other.

What a diagnosis changes

Does knowing Charlie's variant change anything? For treatment, I do not know yet. From what I have read so far, precision medicine programs for this gene are in the works and still early. There may be a dietary supplement he can try, and there may be a clinical trial. A trial designed for Charlie's condition is what I have wanted for 16 years, in place of trying another drug because it sometimes works for other children. I have also wanted a way to anticipate which other organ systems might be affected, and now we have a chance at both. Charlie is 16, he has been through a great deal, and a lot of damage has been done, so I cannot say how much will change for him medically.

Where I am confident a diagnosis will help is in getting services. Until now, Charlie's history read as infantile spasms, which he no longer has because he is no longer an infant, followed by severe seizures and an intellectual and developmental disability (IDD) with no named cause. I had no literature to point to that shows how severe and how lasting his condition is. Other families may be surprised by how much proof is demanded, even for a child who cannot speak and has had multiple epilepsy surgeries. I am paying thousands of dollars out of pocket for a private neuropsychological evaluation, because some of the people willing to assess him for free have raised possibilities as unrealistic as driving a car. My impression is that programs sometimes try to fit a child into a vocational track they already offer, and acknowledging how impaired he is would rule that track out.

A named rare disease, with published papers describing children who are as severely affected as he is, gives me something to show a reviewer. It is a constant battle even with that kind of evidence, and it has been far worse without it.

Over the years I have often been asked what a diagnosis would really change, and whether it might be better not to know. My answer is to let the family decide that. If Charlie qualifies for an aide because of this diagnosis, I can go to my other children's soccer games, Charlie can come along with hands-on support and have a good time, and I can sit with one of his siblings without worrying about catching the signs of a seizure. That changes life for everyone in our family, and it is a benefit we rarely talk about.

These diseases can leave children needing support 24 hours a day, and families often feel that nothing they do is enough to prove their level of need. The experts who understand how severe these conditions are do not review our applications at school or in the community. The people who do review them need to be educated, and it is close to impossible to educate someone about a condition that has no name.

Easier access to genetic testing

We need to make genetic testing easier to get. The epilepsy community has seen a boom in genetic findings because testing is now far more routine for people with epilepsy, and it was not always that way. Before Charlie's first brain surgery, I remember fighting the insurance company hard, together with his genetic counselor and his neurologist, for a basic panel. We have come a long way since then, in large part because of advocacy, including work I was involved with, and I am proud of that. I see no reason other disease areas cannot do the same.

The families who are still searching

My first instinct after learning Charlie's variant was to give all of my time and energy to the SLC35A2 community. I will get involved with that community, and I also think we have to remember the families who are left behind without an answer. That is the ethical thing to do, and it is also sound science. For years, the community around Charlie's gene held conferences and did its work without knowing we existed, which means it went without Charlie as a member and without my help or perspective. Other people who belong in a gene community may still be out in the unknown, as we were.

We also have more in common than we have differences. Seizures and intellectual disability are shared across many of our conditions, and each condition then has its own distinguishing features. Helping families who are still undiagnosed and working on what we share is how a rising tide lifts all boats.

Working across diseases

Those of us who have been through this should reach across disease areas, to the extent we have the energy. Many people with ALS are in the position we were in, with no known cause for their disease. If the epilepsy community has worked out strategies that help, we should bring them to ALS. ALS is where I work and it is my passion, and I need more help there.

I often think about the ALS doctors and the epilepsy doctors who sit next to each other on the same floor of the same building. They may share a data engineer and a biomedical informatics department, yet they run separate studies with them, apply for separate grants, and sometimes compete for the same ones. That cannot be the most efficient arrangement. I would like us to think in terms of neurology informatics, in place of an epilepsy informatics project and an ALS informatics project. That is how I work: I help anyone who comes to me, whatever their disease, and we all learn from one another that way.

I am sure I will have more thoughts as time goes on. One day in, what I see most clearly is how much opportunity we have to change the world. This is Charlie’s legacy. Sixteen years is a very, very long time.