Non-natural peptides containing  electron-rich aromatic subunits have demonstrated the  remarkable  ability  to  spontaneously  assemble  into  long  fibers  with  optical  and electronic responses similar to conventional silicon electronics. These molecules have the  potential  to  serve  as  new  biocompatible  organic  electronics  with  uses  in medical interventions and clean energy.

The  space  of  possible  molecules  is  too  vast  to  search  by  trial-and-improvement experiment, and data-driven modeling techniques can greatly accelerate the search for promising  peptide  designs.  Computationally,  we  have  (i) developed an inexpensive molecular  model  assembly  to  identify  promising  new  chemistries  for experimental testing, and (ii) used data-driven modeling to predict which peptide properties  tend  to  lead  to  desired  structures.  Experimentally,  we  have  used  the insights  from  this  data-driven  modeling  and  have  prepared  peptide  sequences expected  to  exhibit  high  extents  of  cofacial  pi-stacking.  The  electronic  elucidation  of these interactions is ongoing.