CellWorld™

  • Technology Plan

There is more to the human genome project than drug therapies. The CellWorld™ strategy is to develop four new stand-alone technologies, that when combined create a new biomedia experience. The common theme is communication.

The four enabling technologies are:

Technology 1: Knowledge Mapping

Technology 2: Animating Biological Pathways

Technology 3: RenderBank Biomolecular Ray Tracing

Technology 4: Gene Navigator: Interactive Chromosome Rendering & Sonics


Technology 1: Knowledge Mapping Shows The Answer

Our long term objective is to create cellscapes for CellWorld that are factual, detailed and vivid. According to Richard Drake, Ph. D. Biochemist at UAMS, "The cure to cancer has already been discovered, we just don't know it yet." Too much information in too many places. The past two decades of computer science teaches that the best approach to solving new problems is a combined approach.

Naive data-mining and text analysis produces, cluttered, blurred and indefinite results. One cannot throw English text at a machine and expect crisp and clear reasoning to emerge. Fact extraction requires an interactive participation of human experts. Knowledge mapping provides the informational road map required to navigate the incredibly rich space of genetic and biological knowledge.


There are many ways to graph things, but only a few effective ones. Years of experience have shown us the visual representations that work best.

Remember: KM Shows the Answer

Applications: Gene hunting. Cell Process Mapping


Deployment: This will become a 3D graphical search engine, users will list with the search engine by providing content mapped using tools we provide, compatible with Netscape and Explorer, Windows and MacOS.

Startup Requirements: 4 developers + 1 technical director
Startup Timeframe: Winter 2001
Market: Everyone Who Uses Biotechnology Knowledge


Technology 2: Animated Biochemical, Genetic and Signaling Pathways

The world's knowledge about the fate of biomolecules is represented as complex printed schematics. The difficulty with these metabolic blueprints is that they consume a large amount of visual real-estate, they are static, and they require manual reading to understand. This manual reading is time-intensive and results in an understanding limited by persistence of memory .

These three problems can be solved by animating the fate of specific substances. Animations can produce a sixteen fold reduction in visual real-estate requirements while simultaneously eliminating the need for manual reading. Mere observation of these animations transmits in seconds, understanding that formerly required weeks of manual reading. A famous SIGGRAPH animation showed the tomato bush stunt virus in all the splendor of its icosahedral geometry, unfolding one section at at time. In 3 minutes, 2000 people knew the same structural relationships that took a group of scientists 20 years to uncover. This is leverage.

This has also been demonstrated by examples created by WDV, such as glycolysis pathway, isoprenoid biosynthesis pathway and others.

There are three kinds of pathways:
Metabolic Pathways
Gene Expression Pathways
Signal Transduction Pathways

Applications: Cellular Simulation, Disease Understanding
Product: A set of interactive and recombinable animations.
Approach: Off the Shelf Tooling combined with
Off the Shelf Data to produce QuickTime movies and modular molecular datasets.

Startup Requirements: 4 developers + 1 technical director
Startup Timeframe: Winter 2001
Market: Biotechnology, Medicine, Education


Technology 3: RenderBank Biomolecular Ray Tracing

Viewing Planet Cell

According to Moore's Law, "Computer price-performance doubles every 18 months." That means one of two things:

1) the performance of a computer of a given cost doubles or
2) the cost of a computer of a given performance
halves

and this takes place every year and a half.

A factor of two is noticable, but the accumulation of performance is spectacular. In fifteen years, ten doublings produces a speed-up (and/or cost-down) of over a thousand (2^10 = 1024.)

This exponential growth in computer capability is enabling new approaches to image generation and visual realism.

We propose the creation of a "RenderBank", a set of 1024 Gigahertz class workstations that will be used to provide real time ray tracing of complex molecular models for direct production work. The image generation of ray tracing is perfect for the virtual worlds that must be created to convey the cellscapes of CellWorld™. The widespread availability of public domain rendering software and the falling cost of computers makes this a practical solution.

RenderBank will be used in three modes:

Mode One: Real Time Molecular Modeling for CellWorld™ Scene Construction

Mode Two: High Resolution Scene Rendering for 35mm Movie Production

Mode Three: System Rental to Biotechnology Investigators

Startup Requirements: 4 developers + 1 technical director
Startup Timeframe: Winter 2001
Market: Biotechnology, Medicine, Education
, Entertainment

RenderBank Phase In Plan:


Generation Processors Date Staff Cost Task
0 1 Sep-2000 1 $1,500 develop brokers
1 2 Oct-2000 2 $2,622 test CORBA brokers
2 4 Dec-2000 4 $4,582 scale RenderBank
3 8 Jan-2001 4 $8,009 "
4 16 Feb-2001 6 $13,998 "
5 32 Mar-2001 6 $24,467 "
6 64 Apr-2001 6 $42,763 "
7 128 May-2001 8 $74,742 "
8 256 Jun-2001 8 $130,636 "
9 512 Jul-2001 8 $228,328 "
10 1024 Aug-2001 10 $399,076 scale RenderBank
* Incurred Cost Assumes Moore's Law Remains In Force (true since 1978)     

The system will be grown in powers of two system to minimize integration costs and exploit Moore's Law during phase-in.

An example animation has already been developed using this technique.


Technology 4: Gene Navigator: Interactive Chromosome Rendering & Sonics

Viewing the Blueprints

As large DNA sequences become available, new interactive tools will be required that enable chromosomes to be drawn and heard in novel and imaginative ways.

Thirty five percent of the human genome consists of repeated, nonfunctional sequences called ALU repeats. An interactive chromosome drawing program would abbreviate and compress the drawing where these repeats occurred by substituting special colored symbols. The consecutive identification and illustration of the genome using such symbols creates a valuable asset to genome explorers.

An additional sixty percent of the human genome does not code for protein, but consists of "intronic" sequences. These intronic sequences can play a role in gene expression but are not themselves expressed. Many of these introns consist of short, long and tandemly repeated elements known as SINE and LINES. Understanding the frequency, type and position of SINES and LINES gives us insight into chromosome structure, and enables comparisons of genes by individual, origin, heritage, and species. Replacing long repetitive sequences with abbreviated symbols enables a comparative "at-a-glance" understanding of the genetic code.

The remaining three to five percent of the genome codes for proteins that make us who we are. It is essential to understand the placement, relationship and variation that characterize these genes. Some are repeated, some are not. Cancer susceptibility depends on the presence of multiple copies of genes that code to process toxins such as nitrosamines frequently encountered in the diet. Individuals with fewer copies would benefit from seeing the output of their gene chip arrays meaningfully rendered.

The genome is a one dimensional stream. The creation of useful sonic mappings for understanding the genetic territory is an effective communication tool, since sound itself is a one dimensional stream.


Startup Requirements: 4 developers + 1 technical director
Startup Timeframe: Winter 2001
Market: Biotech, Medicine, Media Companies, Entertainment


Summary:

Four stand-alone technologies have been proposed that enable vivid new views of molecular biology, the cell and the genetic codes. The choice of these technologies has not come by accident. Each plays a critical role in the creation of a motion picture called CellWorld™, which will accurately portray the functioning of the cell in a way never before experienced.


Acknowledgements:

Appreciation is expressed to a great legend in computer graphics, Russ Fish of the University of Utah Computer Graphics Lab, who was kind enough to review RenderBank portion of this proposal. The kind advice of Rod Bogart of Industrial Light and Magic is also gratefully acknowedged. Richard Drake at UAMS has been an inspiration, as have a multitude of biochemists who have given their lives and careers to bring us to the magic of today.