Information Processing in Chemical Networks (Part 1)

There’s a workshop this summer:

Dynamics, Thermodynamics and Information Processing in Chemical Networks, 13-16 June 2017, Complex Systems and Statistical Mechanics Group, University of Luxembourg. Organized by Massimiliano Esposito and Matteo Polettini.

They write, “The idea of the workshop is to bring in contact a small number of high-profile research groups working at the frontier between physics and biochemistry, with particular emphasis on the role of Chemical Networks.”

The speakers may include John Baez, Sophie de Buyl, Massimiliano Esposito, Arren Bar-Even, Christoff Flamm, Ronan Fleming, Christian Gaspard, Daniel Merkle, Philippe Nge, Thomas Ouldridge, Luca Peliti, Matteo Polettini, Hong Qian, Stefan Schuster, Alexander Skupin, Pieter Rein ten Wolde. I believe attendance is by invitation only, so I’ll endeavor to make some of the ideas presented available here at this blog.

Some of the people involved

I’m looking forward to this, in part because there will be a mix of speakers I’ve met, speakers I know but haven’t met, and speakers I don’t know yet. I feel like reminiscing a bit, and I hope you’ll forgive me these reminiscences, since if you try the links you’ll get an introduction to the interface between computation and chemical reaction networks.

In part 25 of the network theory series here, I imagined an arbitrary chemical reaction network and said:

We could try to use these reactions to build a ‘chemical computer’. But how powerful can such a computer be? I don’t know the answer.

Luca Cardelli answered my question in part 26. This was just my first introduction to the wonderful world of chemical computing. Erik Winfree has a DNA and Natural Algorithms Group at Caltech, practically next door to Riverside, and the people there do a lot of great work on this subject. David Soloveichik, now at U. T. Austin, is an alumnus of this group.

In 2014 I met all three of these folks, and many other cool people working on these theme, at a workshop I tried to summarize here:

Programming with chemical reaction networks, Azimuth, 23 March 2014.

The computational power of chemical reaction networks, 10 June 2014.

Chemical reaction network talks, 26 June 2014.

I met Matteo Polettini about a year later, at a really big workshop on chemical reaction networks run by Elisenda Feliu and Carsten Wiuf:

Trends in reaction network theory (part 1), Azimuth, 27 January 2015.

Trends in reaction network theory (part 2), Azimuth, 1 July 2015.

Polettini has his own blog, very much worth visiting. For example, you can see his view of the same workshop here:

• Matteo Polettini, Mathematical trends in reaction network theory: part 1 and part 2, Out of Equilibrium, 1 July 2015.

Finally, I met Massimiliano Esposito and Christoph Flamm recently at the Santa Fe Institute, at a workshop summarized here:

Information processing and biology, Azimuth, 7 November 2016.

So, I’ve gradually become educated in this area, and I hope that by June I’ll be ready to say something interesting about the semantics of chemical reaction networks. Blake Pollard and I are writing a paper about this now.

8 Responses to Information Processing in Chemical Networks (Part 1)

  1. Gary Lewis says:

    Not sure what you mean by the semantics of chemical reaction networks. To me it conjures up an image of category theory applied to metabolic pathways, which are conveniently near universal for all species. This would pick up on the work of Harold Morowitz that goes back decades, and the very recent publication of a book by Smith and Morowitz (The Origin and Nature of Life on Earth). Am I anywhere near to what you meant? … Regards, Gary

    • John Baez says:

      I wish my work were closer to the biological ideas you’re talking about! It may take a while to get there. There are many kinds of ‘meaning’ for a chemical reaction network, and each gives a semantics that one can try to formalize. I’m just getting started on this. Here you can see a talk I gave on functorial semantics applied to chemical reaction networks (or actually Petri nets, which are another way of drawing the exact same information):

      • John Baez, Compositionality in network theory.

      You can see the slides here, with links to papers, and a video here:

      • Ken Spicer says:

        I’m actually fascinated by this sort of thing. I’ve been doing logic circuits for decades and my first question I asked my brother (who is a computer scientist by trade) is why we have to operate in binary. Found out it’s because electricity just seems to work best in on or off states and bumping up the number system just increases complexities without gains.

        I recently saw an article talking about the sequencing of DNA to store tons of data and retrieve it without errors but the one thing no one was talking about was read/write speed. I’m curious how slow it currently is, because in my head someone is sitting there with a dish and a list of letters for the years since they discovered DNA.

        • John Baez says:

          I too heard that recent news about DNA data storage, but I’m always suspicious of such claims in popular media because they leave out a lot of details (like how fast the storage and retrieval are), emphasize the good news, and downplay the bad news. Could you read the original paper and report back to us?

          • Yaniv Erlich and Dina Zielinski, DNA Fountain enables a robust and efficient storage architecture, Science, 3 March 2017.

        • Ken Spicer says:

          Sure can try!

          The paper itself directly references the mathematical method they seem to be using and naming it after, namely fountain codes! ‘DNA Fountain’ uses a Luby transform (or LT codes) which are useful for this particular type of encoding because DNA seems to be a bit lossy by nature, but you can correct for errors by using the same sort of of one-way error correction used by cell-phones and mobile TV and can reliably transmit data even when subject to erasure.

          As expected, their bits are larger than the regular bit! It’s actually four characters directly mapped to binary, {00,01,10,11} to {A,C,G,T} respectively. If everything were perfect, each nucleotide could theoretically hold 2 bits (4 choices) but they run into a number of limitations with biochemistry including some sorts of sequences being undesirable due to difficulty to synthesize and error proneness which reduces the actual coding potential of each nucleotide down to ~1.98 bits (through quantitative analysis), oligonucleotide synthesis (the bits of bits they’re making), PCR, and even just storage results in to an overall Shannon information capacity of ~1.83 bits per nucleotide as their upper bound. Exploration of this seems to be one of the main points of the paper, as previous DNA storage studies only realized half of this Shannon information capacity and had issues with perfect data retrieval.

          Other key points:

          1) They were able to to achieve an information density of 1.57 bits/nucleotide.

          2) Using their method, you can reach file sizes of up to 500Mbytes

          3) They generated code that contained 72,000 oligos with a 7% redundancy, this number was chosen due to cost efficiency with the manufacturer that actually sequenced the DNA. Meaning even they don’t reference how slow writing actually is, turn around could potentially be months.

          4) Decoding a 2.1MB file takes them 9 minutes which makes it have a read speed of about 3.88kBps

          5) The cost was $3500/Mbyte per write in this particular study. Meaning that particular file probably cost closer to $7000 to write.

        • John Baez says:

          Thanks! That’s very interesting. It may be that DNA storage will only be good for storing large amounts of data for long periods of time, e.g. keeping a history of our civilization in the Svalbard Global Seed Vault, shown here:


          Of course, for this one would need to investigate how long information stored in DNA at low temperatures can last.

  2. John Baez says:

    There’s now a website for this conference:

    Dynamics, Thermodynamics and Information Processing in Chemical Networks, 13-16 June 2017, Complex Systems and Statistical Mechanics Group, University of Luxembourg. Organized by Massimiliano Esposito and Matteo Polettini.

  3. I’m in Luxembourg, and I’ll be blogging a bit about this workshop:

    Dynamics, Thermodynamics and Information Processing in Chemical Networks, 13-16 June 2017, Complex Systems and Statistical Mechanics Group, University of Luxembourg. Organized by Massimiliano Esposito and Matteo Polettini.

    I’ll do it in the comments!

You can use Markdown or HTML in your comments. You can also use LaTeX, like this: $latex E = m c^2 $. The word 'latex' comes right after the first dollar sign, with a space after it.

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s