Clarification of Time Course Phenotype

[VarenyaJ]

Hi all,

I wanted to ask a question regarding how time course tracking of phenotypic observations can or should be implemented in Phenopackets.

---

1. Evolving phenotype example

For instance, suppose a physician diagnoses a fetus with Decreased head circumference (HP:0040195) at a GestationalAge of 14 weeks, but at a later scan (20 weeks) the sonographer finds that the head circumference is below the 3rd percentile and instead documents Microcephaly (HP:0000252).

In this type of case, would this change over time be represented through:

• the Update class,
• the TimeElement class,
• the onset property,
  or by some other mechanism intended for longitudinal data?

The key question here is how to encode that a phenotypic feature evolves or progresses between two time points rather than representing two entirely separate features.

---

2. Recurring phenotype example

Similarly, how would this apply to a recurring phenotype such as Seizure (HP:0001250)?
In this case, a patient may experience multiple seizure events over time, each with a distinct onset, duration, or context. Would each seizure episode be represented as a separate PhenotypicFeature, or is there a recommended structure for grouping repeated occurrences of the same feature?

---

3. Compound phenotype with interventions

Finally, how would this approach extend to compound or structural conditions such as Tetralogy of Fallot (HP:0001636)?
As I understand it, this condition is defined by a combination of:

• Pulmonary stenosis -- narrowing of the right ventricular outflow tract,
• Ventricular septal defect -- a hole between the two ventricles,
• Right ventricular hypertrophy, and
• Overriding aorta -- where the aorta receives blood from both ventricles.

I'm wondering how the schema should represent the temporal course of such a composite condition -- for example, if the patient undergoes a reconstructive or transplant surgery to repair an atrioventricular canal defect (Tetralogy of Fallot with atrioventricular canal defect, HP:0011677), but later presents again with residual or new issues such as pulmonary atresia (Tetralogy of Fallot with pulmonary atresia, HP:0012516).

Would this kind of longitudinal progression or reclassification be expressed through Update events, separate PhenotypicFeature instances linked by a TimeElement, or another construct?

---

Summary question

Overall, what is the recommended or intended approach in Phenopackets to capture:

• evolving phenotypic features over time,
• repeated or recurrent observations of the same phenotype, and
• compound conditions that are modified by clinical intervention or reclassification?

I'd greatly appreciate any examples or guidance on best practices for representing temporal and longitudinal phenotypic data within the Phenopacket schema.

Thank you for your time and insights!