SD2: Synergistic Discovery and Design

Background:

Synergistic Discovery and Design (SD2) aims to accelerate the Design Build Test Learn (DBTL) loop for laboratory science. SD2 focuses upon synthetic biology (genetic circuits), protein synthesis, and perovskite crystal formation. SIFT’s XPlan software makes notional experiments actionable by configuring experimental platforms to produce informative measurements of experimental constructs.

SIFT Approach:

SIFT’s contributions to SD2 include not only the XPlan experimental planner, but also several related analysis tools. In order for XPlan to design informative experiments, we developed analyses that both predict the outcome of experiments and process actual experimental results. Our circuit prediction tool models genetic circuits and the experimental conditions as a hierarchial probabilistic model. It learns from experimental data to parameterize the model and predicts not only the circuit behavior but also the anticipated information gain achievable with additional samples. We created analyses to compute sample growth rates and predict the number of live cells in an experimental sample. These analyses help isolate failed experiments and improve the measurement quality of successful experiments. With these innovations, XPlan automates high-throughput screening and experiment design for the test and learn segments of the DBTL loop.

Benefits:

  • Can predict the outcome of experiments and process actual experimental results
  • Can isolate failed experiments and improve the measurement quality of successful experiments
  • Automates high-throughput screening and experiment design for the test and learn segments of the DBTL loop