I'd argue it simply makes them a bad data scientist. Biology cares not for the categories we create to explain it, and the purpose of categorization is to make sense of what's already in the world, not to prescribe how it should be. Exceptions exist everywhere, not just in trans people. If your modeling of the data is inaccurate because you only have a binary categorization of sex, that categorization is to blame, not the people who the data represents.

So ultimately, in medical studies, perhaps it's important to note how you categorized your subjects' sex, how that relates to the mechanisms of what you're studying, and perhaps studying trans people's data further can provide more insights e.g. how hormones affect a condition. Science and data is reliant on the narratives we use to inspect and describe it, and the less of our societal baggage we impose on that process, the better.