.png)
.jpg)
Drug discovery fails late when uncertainty isn’t addressed early.
In this episode of Data in Biotech, Ross Katz speaks with Brant Peterson, Vice President & Fellow, Data Science at Valo Health, about what it really means to “start with the end” in modern drug discovery, and how patient data, genetics, and causal reasoning shape decisions long before a molecule reaches the clinic.
They explore:
>> Why phenotype, not models or targets, should anchor discovery decisions
>> How real-world data and genetics help de-risk biology earlier than traditional workflows
>> Where causal DAGs, pleiotropy, and hypothesis prioritization actually matter in practice
>> Why understanding data generation and bias comes before any machine learning model
>> How patient subgroups, not average patients, reshape thinking in neurodegenerative disease
The conversation reframes data science in biotech as less about prediction and more about mechanistic clarity: deciding which hypotheses are worth pursuing, which evidence belongs where, and which paths can realistically lead to a medicine.
Listen to the full episode of Data in Biotech:
.jpg)
.jpg)
.jpg)
