Platforms for Biosensing and Analytics
Adapting electronic technologies to access the molecular world
Under active Development
Biosemiotics & Language Discovery
Applying machine learning to biological systems to accelerate understanding
Biosemiotics takes the view that biology is embodied as a set of interacting languages; the Central Dogma of DNA-RNA-Protein is an example of coding, with peptide sequences generating motifs in 3D structures and their emergent properties.
Perhaps the most limiting aspect to scaling up synthetic biology is our nascent understanding of complete biological systems, making it more difficult to design and predict how changes will affect a metabolic network, organism, or ecosystem. Without clear design principles, discovery can be a slow process, fraught with pursuing dead ends or impractical solutions.
Machine learning has started to transform and accelerate discovery. A recent Nobel Prize was awarded based on using AI to more accurately predict protein folding, a notoriously difficult computational problem using other techniques. Large Language Models show promise in exposing the patterns in biological data.
We use this perspective to drive multi-modal LLMs, trained on public and private datasets, to generate new insights and models about the underlying languages of life. Anaphasic is working on projects in this space to first test regeneration or “rediscovery” of known biological principles, to extend these into areas which haven’t been fully explored, and ultimately, to discover potential new linguistic connections between different aspects of biological systems.
More information will be available as this project develops
DNA Data Storage
Using a three-billion year old technology to enable and preserve our digital future
Today, the world is creating data at a much faster rate than storage technologies can handle. There is a significant risk that within two decades, buying exponentially more storage capacity will become prohibitively expensive and resource-constrained.
Compounding this, none of our existing technologies can guarantee access for more than a couple decades, leading to the risk of losing the stored information unless rewritten to new media.
Nature already provides us with a working example for massive data storage that has functioned for billions of years: DNA. The information density is on the order of tens of atoms per bit — ultimately exabytes per gram. No existing technology comes close, because every bit is associated with not only a memory circuit or magnetic domain or optical structure, but all the millions of atoms in the substrate for each. We also have recovered data — genomes — from samples unintentionally stored for tens of thousands of years. By adding intrinsic error correction and engineered materials and conditions, we can retain data with this kind of longevity.
Learn more about the IARPA-funded program that contributed $50M to developing these technologies:
David Markowitz SC22 talk: Biology is all You Need
(https://youtu.be/HGOpHsZDIGU)
A presentation on DNA Data Storage for NerdNight
in San Francisco, July 2023
(local pdf)
The Semiconductor Research Corporation SemiSynbio Industry Roadmap, developed via a grant from NSF. I was part of the Executive Committee and chaired of the topics / chapters:
SSB Industry Roadmap (local pdf)
Platforms for Biosensing and Analytics
Adapting electronic technologies to access the molecular world
Molecular information is at the core of understanding biology, of life’s structures and processes. We take the perspective that there are two aspects of interacting with molecular information: sensing / actuation, the interface between the physical and digital worlds, and analytics, the processing of molecular information in the digital regime.
Over time, access to molecular information has benefitted from increasing levels of miniaturizing and integration, ultimately using the surface of an electrically active silicon substrate to directly support synthetic chemistry and direct electrical detection of molecular interactions.
Likewise, we have developed increasingly optimized software and hardware to process signals and analyze data as life sciences have become more digital and subject to more conventional modes of engineering.
Indeed the introduction of semiconductors and computing into life sciences has and continues to revolutionize our understanding of living systems. Anaphasic have developed deep expertise in both of these areas, combining them for an overall systems approach to opportunities across life sciences and computing.
Frameworks for Complex Systems
Developing comprehensive, flexible frameworks and methods to discover, design, develop, and deliver successful solutions
Today, the world is creating data at a much faster rate than storage technologies can handle. There is a significant risk that within two decades, buying exponentially more storage capacity will become prohibitively expensive and resource-constrained.
One of the most challenging things about truly complex systems is coming up with ways to organize and structure them to be more comprehensible and manageable without missing critical information. Engineering systems at the limits of complexity has been a focus throughout our history.
For more detail about these frameworks, check out these external resources now, and check back later on this site:
Link to “Balanced Solutions” Book
-https://a.co/d/jf2BT3V
Link to Erik Simmons presentation at Construx
-https://youtu.be/iWOTupPbSYM