A mechanistic model of the BLADE platform predicts performance characteristics of 256 different synthetic DNA recombination circuits

Abstract

Boolean logic and arithmetic through DNA excision (BLADE) is a recently developed platform for implementing inducible and logical control over gene expression in mammalian cells, which has the potential to revolutionise cell engineering for therapeutic applications. This 2-input 2-output platform can implement 256 different logical circuits that exploit the specificity and stability of DNA recombination. Here, we develop the first mechanistic mathematical model of the 2-input BLADE platform based on Cre-and Flp-mediated DNA excision. After calibrating the model on experimental data from two circuits, we demonstrate close agreement between model outputs and data on the other 111 circuits that have so far been experimentally constructed using the 2-input BLADE platform. Model simulations of the remaining 143 circuits that have yet to be tested experimentally predict excellent performance of the 2-input BLADE platform across the range of possible circuits.Circuits from both the tested and untested subsets that perform less well consist of a disproportionally high number of STOP sequences. Model predictions suggested that circuit performance declines with a decrease in recombinase expression and new experimental data was generated that confirms this relationship.

Programmable Mixed-Signal Biocomputers in Mammalian Cells

“…Each TF from the Layer 1 circuit can target up to three unique outputs, specified by the truth table in Figure 4B . Without optimization, ∼90% of circuits perform as expected based on a vector proximity measure ( < 15°, [5] [40]).…”

“…A previously established and refined vector proximity metric was used to determine the performance of all digital circuit operations performed in this work [5], [40]. Briefly, circuit data was broken down into two vectors t and s , where t represents an n -dimensional binary vector corresponding to the intended truth table of an individual circuit and s represents the corresponding signal vector composed of fluorescence output values for each address.…”

“…Briefly, circuit data was broken down into two vectors t and s , where t represents an n -dimensional binary vector corresponding to the intended truth table of an individual circuit and s represents the corresponding signal vector composed of fluorescence output values for each address. Recent work by Bowyer et al has shown that the original vector proximity measure can be improved by dividing the original vector proximity metric by the number of fluorescence-expressing addresses ( n f ) [40], as it is more difficult for transfected cells to achieve a full population response for fluorescent states than it is for all cells to remain inactive in non-fluorescing states. This modified cosine similarity metric between vectors t and s is calculated as where and …”

Synthetic memory circuits for programmable cell reconfiguration in plants

“…In bacteria, yeasts, and mammalian cells, transcription factors and recombinases have been used to construct gene circuits in cell culture [12][13][14][15][16][17][18][19][20][21][22][23] . However, in plants, only a limited number of circuits have been developed to date.…”