Random Lessons from Nano-bio
for NSF Materials Science
Eric Jakobsson
University of Illinois at Urbana-Champaign
Director, NIH Roadmap National Center for
the Design of Biomimetic Nanoconductors
Founder, Biology Student Workbench
Former Director, NIGMS Center for
Bioinformatics and Computational Biology,
Former Chair, NIH Biomedical Information
Science and Technology Initiative Consortium
Google: Nanoconductors—1st hit
Official Organization Chart of the University of Washington—a typical official
research university organization--guaranteed to fail if functional relationships
followed the chart. The system only works at all because people create other
connections not reflected in the official chart, but the system works poorly
because the institution does not reward the creation of the other connections.
An organization chart that needs work
Organization charts that actually work (from Tong et al,
Global mapping of the yeast genetic interaction
network, Science, 2-6-04…). Typically, there are
multiple paths between each pair of nodes.
An organization chart that we think is working---the interaction network of
participating investigators at The National Center for the Design of Biomimetic
Nanoconductors. The Principal Investigator (Jakobsson) is at the center rather than
the top. There are multiple alternative pathways to getting things done. It looks a
lot more like the gene interaction network than a typical research institution,
Bridging the Experimental-ComputationalDesign Divides—General Principles
• Theory and computation in the Center will
be validated by experiment in the Center.
• Experiment in the Center will be guided by
theory and computation in the Center.
• Computation and theory in the Center will
be focused on support of device design.
Bridging the Experimental-ComputationalDesign Divides—Specific Example 1
• We would like to fabricate nano/micro scale “cells” with a
self-assembled membrane supported by semipermeable
(permeable to water but not ions) silica or other thin
films---essentially fabricated aquaporins. We know from
electrostatics theory that this should be possible by
adjusting the image force barrier in nanopores.
• Brinker lab (experimentalists) have devised an
electrolytic deposition technique that enables him to
closely control the size of nanopores in a thin film.
• Aluru lab (computationalists) are doing molecular
dynamics simulations to determine range of nanopore
dimensions for the materials from the Brinker lab that will
be semipermeable; i.e., pass water but not ions.
Bridging the Experimental-ComputationalDesign Divides—Specific Example 2
• We want to design robust ion channels from
beta-barrels that will have properties of our
specification for membrane-based devices we
• Bayley lab (experimentalists) have developed
the ability to insert cyclodextrin into alphahemolysin pore, in a position where it could act
as a selectivity filter.
• Roux lab (computationalists) are doing free
energy/molecular dynamics simulations to
determine selectivity of chemically decorated
cyclodextrins to be synthesized in Bayley lab
Bridging the Experimental-ComputationalDesign Divides—Specific Example 3
• We would like to design and fabricate nanoscale
batteries from supported membranes—biomimesis
based on the organ of the electric eel as a proof of
• Jakobsson lab (computationalists) developed a dynamic
model describing ion and water flow across the airway
• LaVan lab (device design/experimentalists) have
adapted the epithelial model as a template to do a first
draft model of the electric organ, as a foundation for
nano-battery design.
• Plimpton lab (computationalists) are building the LaVan
first draft model into a full 3-dimensional model of the
electric organ.
Bridging time and length scales in understanding
membrane dynamics and organization
• Problem: No supercomputer that we can foresee will be able to
simulate, from atomistic molecular dynamics, domain formation and
phase relationships in heterogeneous membranes
• Our labs’ (Scott/Grama/Jakobsson) approach is to use atomistic
molecular dynamics simulations to parameterize Mean Field
Langevin Dynamics simulations that can span large distances and
long time scales.
• In our specific implementation, cholesterol molecules are discrete
particles, while other membrane lipids are represented by
continuous fields of concentration and order parameter. Field
evolution dynamics are derived directly from analysis of molecular
dynamics output of corresponding membrane, specifically from
correlation analysis of neighbor interactions.
• Method has been validated by successfully reproducing heat
capacity through phase transitions, and phase boundary tie lines, for
dppc-cholesterol mixtures. (ms. under review)
Suggested dissemination guidelines for software developed
under CI (adapted from NIH guidelines and taking account
of discussion at NSF workshop)
• 1) The software should be freely (or at very nominal cost) available
to researchers and educators in the non-profit sector, such as
institutions of education, research institutes, and government
• 2) The terms of software availability should permit the
commercialization of enhanced or customized versions of the
software, or incorporation of the software or pieces of it into other
software packages.
• 3) The terms of software availability should include the ability of
researchers outside the supported project to modify the source code
and to share modifications with other colleagues, with obligation to
share modifications also with original developers.
• 4) The software must be in a form such that if the development team
loses interest in the software subsequent to the life of the project
another individual or team can make use of previous work to
continue development and maintenance.
User Input
User Web Browser
Output to User
and queries
(May have varying
interfaces and be
written in different
to User
Format Translator,
(May be of
Query Engine and
varying formats
Program Driver
A question that emerged at a biology education
meeting in North Carolina
Could homo sapiens interbreed, or
have interbred, with other species?
Bottom Lines
• NSF CI (and all research institutions) should consider reorganization
along lines proved to work in biological interaction networks and
suggested by leading edge social network theory
• CI should nurture interactions among experimentalists,
computationalists, theoreticians, and designers.
• CI should nurture development, implementation, and dissemination
of efficient multiscale.
• NSF CI should adopt software dissemination guidelines similar to
NIH roadmap BISTI National Centers for Biomedical Computing.
• NSF CI should promote portal development that provides unified
access to multiple data sources and computational analysis tools for
research and problem-based education. Translation between data
and interface formats should be done by the portal developers.

Random Lessons from Nano-bio for NSF Materials Science