Data

Tools

Indexes

Applications

Experiments

Awesome

An open API service indexing awesome lists of open source software.

https://github.com/husonlab/catrenet

Autocatalytic networks
https://github.com/husonlab/catrenet

Last synced: 19 days ago
JSON representation

Autocatalytic networks

Awesome Lists containing this project

README 

          
# CatReNet



Splash



CatReNet (formerly CatlyNet) is a program for working with (auto-)catalytic reaction networks, based on a set of

'catalyzed reactions' and a given 'food set' of molecules  [[HXS24b]](#70).

It provides fast and exact algorithms for calculating three main types of autocatalytic networks,

namely (maximal) RAFs, CAFs, and pseudo-RAFs. These three notions are formalizations of the concept of a 'collectively

autocatalytic set' originally introduced by Stuart Kauffman. Several other algorithms are also implemented.

The program can visualize a CRS and animate the emergence of a maximal auto-catalytic network.

CatReNet was first introduced as CatlyNet in [[SXH20]](#25).



CatReNet is written in Java using JavaFX and

is loosely based on a web application written by Dietrich Radel: http://www.math.canterbury.ac.nz/bio/RAF/.

This software was developed by Daniel H. Huson, Joana C. Xavier and Mike A. Steel, see [[HXS24b]](#70).



## 📦 Installers



Interactive installers for MacOS, Windows and Linux are

available [here](https://software-ab.cs.uni-tuebingen.de/download/catrenet/welcome.html or [here](https://unitc-my.sharepoint.com/:f:/g/personal/iijhu01_cloud_uni-tuebingen_de/EmvzMdVieMZMijtxKhVbI-oBqq9fGYHKhVnW4vqLlZ3_IA?e=U2YIwh).



Versions for Android and iOS are under development.



## 🎥 Video Introduction



We’ve prepared a short video on how to get started using CatReNet:



👉 [Watch on YouTube](https://youtu.be/8sBZtX101gU)



## 🚀 Getting started



Install the software using the interactive installer.



The installation directory contains a directory of examples. Select the `File->Open` menu item, navigate to the examples directory and select an example file to open:



image.



Then select an algorithm to run:



image



Or view the network:



image



## ✨ Main features



Input to CatReNet is a catalytic reaction system (CRS), which consists of a list of catalyzed reactions and a food set

of molecules, specified like this example consisting of two food items and three (one-way) reactions:



```

Food: f1 f2



r1 : f2 [f1,p3] => p1

r2 : p1 [f2] => p2

r3 : p2 [p1] => p3

```



Several algorithms are available that aim at computing certain 'auto-catalytic' subsystems.

The three main calculations are the computation of a "maxRAF"

(maximal RAF, reflexively autocatalytic f-generated system), a maxCAF (maximal CAF, constructively autocatalytic and

f-generated system)

and a maxPseudoRAF (maximal pseudo RAF) of a CRS.



The program can display a CRS as a network, in several ways. Moreover, it can animate the emergence of a Max RAF, Max

CAF

or Max Pseudo RAF.



## 🧠 Algorithms and heuristics



In more detail, the program provides the following calculations:



- Max RAF Algorithm - computes the maximal RAF [[HMS15]](#10) (see also  [[H23]](#40))

- Max CAF Algorithm - computes the maximal CAF [[HMS15]](#10)

- Max Pseudo RAF Algorithm - computes the maximal Pseudo RAF [[HMS15]](#10)

- Strictly Autocatalytic Max RAF Algorithm - computes a Max RAF that has the additional property that any contained

  reaction requires at least one molecule type for catalyzation that is not in the food set [[HXS24]](#60)

- Min RAF-Generating Given Element Algorithm - Identifies a subset of the Max RAF that is (i) a RAF and (ii) generates a

  given element x

  (not in the food set) and (iii) which is minimal amongst all such sets satisfying (i) and (ii).

- Trivial CAFs Algorithm - computes all reactions that can run using only the food set

- Trivial RAFs Algorithm - computes all reactions that can run using only the food set, where the catalyst need not be

  in the food set if the reaction produces it

- Core RAF Algorithm - computes the unique irreducible RAF, if it exists (Section 4.1 of [[SXH20]](#25))

- Quotient RAF Algorithm - computes the Max RAF minus all the reactions from the Max CAF and adds all products of the

  Max CAF to the food set [[SXH20]](#25)

- Remove Trivial RAFs Algorithm - computes CRS that is obtained by removing all trivial RAFs

- Min iRAF Heuristic - searches for a smallest irreducible RAF in a heuristic fashion [[HXS24b]](#50)

- MU CAF Algorithm - computes one maximal uninhibited CAF

- U RAF Algorithm - computes a max RAF, removes inhibited reactions, and then recomputes the max RAF

  (Section 6 of [[HMS16]](#20))

- Run MU CAF Multiple Times... - Runs the MU CAF algorithm multiple times, using different orderings of the input

  reactions

- Determine Necessarily Spontaneous in RAF - determine those reactions that must initially run uncatalyzed and then

  beome catalyzed later

- Greedily Grow MaxCAF to MaxRAF - greedily grow maxCAF to maxRAF by making reactions spontaneous

- Compute Reaction Dependencies - computes the graph of dependencies between all food-set generated

  reactions [[HXS24]](#60)

- Compute Molecule Dependencies - computes the graph of dependencies between all molecules [[HXS24]](#60)

- Stratify reactions and molecules - list reactions and molecules ranked by order in which they first appear

- Compute Importance - computes the percent difference between model size and model size without given food

  item [[HS23]](#50)



In addition, the program provides an implementation of Kauffman's  'binary polymer model'.

This features generates a system consisting of all polymers (over a finite alphabet)

of length at most n, with ligation-cleavage reactions, and with catalysis assigned randomly (each molecule catalysing on

average m reactions).



## 📤 Export computed systems



Any of the computed subsystems can be exported to a new file.



## 🌐 Network visualization



The program provides a number or visualizations. Because the visualization of very large graphs can be challenging for

the current

implementation of CatReNet, visualization menu items are disabled for reaction systems containing more than 500

reactions.

You can change this threshold using the settings panel.



Note that the currently implemented network layout algorithm is very basic. For nice layouts,

export a computed network in GML format and then view it in Cytoscape  (see [Export network](#export-network)).



The program can represent a CRS using several types of networks:



- Full network - this shows all reactions, food items, reactants, produces and catalysts as nodes connected by directed

  edges.

- Association Network - reactions are shown as nodes and a directed edge from one reaction to the other indicates that

  the one reaction produces a reactant or catalyst for the other

- Reactant Association Network - reactions are shown as nodes and a directed edge indicates that one reaction produces a

  reactant for the other.

- Reaction-Dependency Network - shows dependencies between reactions (this feature requires testing and debugging)

- Molecule-Dependency Network - show dependencies between molecules (this feature requires testing and debugging)



There are a several options:



- Suppress Formal Food Item - if there is a formal food item present, which is used to implement "catalyst-free"

  reactions, then selecting this option will hide it

- Suppress Catalyst Edges - select this to hide catalyst edges

- Use Multicopy Food Nodes - select this to show each food node multiple times, to produce a less tangled network

- Show/Hide Node Labels - select to hide node labels



## 🧮 Other network calculations



These graph calculations all ignore any catalysts that might be present and are untested on datasets using two-way

reactions:



- Stratification network (Reactions only) - A directed network that aims at capturing the partial ordering of reactions

  based on

  their dependencies: There is an edge from reaction r to reaction s, if reaction r occurs earlier than s in the

  stratification of

  reactions and r has a product that is a reactant for s (and r is the latest reaction to provide that reactant)



- Stratification network (Reactions and Required Molecules) - The same graph as the previous one, except there are now

  edges from r to m to s, where r and s are reactions as described in the previous point, and m is a node that

  represents

  the molecule that is produced by r and used by s. We also have edges m to r, where m is an item of the food set and r

  requires m as a reactant.



- Stratification network (Reactions and All Molecules) - The same graph as the previous one, plus additional nodes

  that represent the other products of reactions.



- Reaction Precedence Network – A directed network that aims at capturing the partial ordering of reactions based on

  their dependencies:

  if an admissible ordering of reactions exists (ignoring catalysts), then this graph is a DAG that represents which

  reactions must occur before others.



## 🎞️ Animation



The program can animate the emergence of three types of systems:



- Animate Max RAF - in this animation, uncatalyzed reactions run at a low rate, as long as all reactants are present.

- Animate Max CAF - reactions only run (at full rate) when catalyzed and all reactants are present.

- Animate Max Pseudo RAF - same as Max RAF, but additionally, molecules that are not being produced can spontaneously

  arise at a low rate.



Run long enough, in each case, the set of reactions running at full rate, together with the assoicated food sources,

constitutes a subsytem of the animated type.



There are several options:



- Animate Inhibitions - if inhibitions are present in the input CRS, then animate them

- Move Labels - during animation, move copies of the molecule-type labels along the edges rather than just disks

- Use Colors - during animation, give different molecules different colors

- Network Embbedder Iterations - set the number of iterations used by the algorithm that computes the layout of the

  network



![Animated GIF](https://github.com/husonlab/catrenet/raw/master/artwork/animation.gif)



## 📎 Export network



The network can be exported as an image in these formats: PNG, SVG and PDF.



Using the `File -> Export -> Network in GML Format...` format menu item, any network can be exported to a file in GML

format

that can be imported into [Cytoscape](https://cytoscape.org) (using that program's `File -> Import -> Network From File`

menu item).



## ⚙️ Settings panel



The settings panel provides a number of configuration panels:



- Notation for reactions: select between full, sparse, and tabbed notation. Select between two different arrow types.

- Default styles for nodes and edges in network: set line style, color and width for all types of edges, and shape,

  color and size of nodes.

- Display labels: Maintains a list of labels to use in network visualizations, for example, in KEGG-based networks,

  `C00005` represents `Reduced nicotinamide adenine dinucleotide phosphate`.

- Animation: decide use different colors for different molecules, decide whether labels are used in animation

- Other settings: maximum number of reactions allowed in network drawing, number of network embedding iterations, and

  whether to wrap text.



## 🛠️ Commandline tools



The Linux and MacOS distributions have a tools directory that contains two commandline programs:



- catrenet-tool - runs the implemented algorithms on one or more CRS input files

- polymer-tool - generates a set of CRS input files using the binary polymer model



## 🗂️ Provided datasets



The program comes with a number of example datasets (from [here](http://www.math.canterbury.ac.nz/bio/RAF/)). They are

placed in a directory called `CatReNet-Examples`

during installation of the program.



- [example-0.crs](examples/example-0.crs) - 6 reactions, 3 food items, has a Max RAF of size 3 and no Max CAF

- [example-1.crs](examples/example-1.crs) - 6 reactions and 12 food itens, is a Max RAF and has no Max CAF

- [example-2.crs](examples/example-2.crs) - uses binary polymers, has 5 reactions and 6 food items, has a Max RAF of size 5 and a Max CAF of size

  3

- [example-3.crs](examples/example-3.crs) - 14 reactions and 1 food item, is a Max RAF and has no Max CAF

- [example-4.crs](examples/example-4.crs) - 9 reactions and 1 food item, is a Max RAF and has no Max CAF

- [example-5.crs](examples/example-5.crs) - 15 reactions and 2 food items, is a Max RAF and has no Max CAF

- [example-6.crs](examples/example-6.crs) - 7 reactions and 1 food item, is a Max RAF and has no Max CAF

- [example-7.crs](examples/example-7.crs) - 3 reactions (1 of which is 2-way) and 4 food items, is a Max RAF and a Max CAF

- [example-8.crs](examples/example-8.crs) - uses binary polymers, has 17 two-way reactions and 4 foot items, is a Max RAF and a Max CAF

- [example-9.crs](examples/example-9.crs) - Has a maxRAF of size 4 (which is an iRAF), a pseudoRAF of size 7 (everything) and no CAF (Figure 2

  from [[XHKSM20]](#35))

- [example-10.crs](examples/example-10.crs) - 3 reactions and 1 food item, is a Max RAF and a Max CAF (Fig. 3(c) in [[GFHKU]](#22))

- [inhibitions-1.crs](examples/inhibitions-1.crs) - 2 reactions (1 of which has an inhibitor) and 6 food items, Max RAF and Max CAF both have one reaction

- [prokaryotic-network.crs](examples/prokaryotic-network.crs) - 6039 reactions and 68 food items. Prokaryotic catalytic reaction network from [[XHKSM20]](#35).



Here are the two examples shown in our getting started video:



- [how-to-build-a-house-maxCAF](examples/how-to-build-a-house-maxRAF.crs) - 6 reactions, 4 food items, has a Max CAF of

  size 6

- [how-to-build-a-house-maxRAF](examples/how-to-build-a-house-maxRAF.crs) - 7 reactions, 4 food items, has a Max RAF of

  size 7 and no Max CAF



## 📚 References



[HSS15]

Hordijk, W., Smith, J.I. and Steel, M.A. (2015). [Algorithms for detecting and analysing autocatalytic sets](https://almob.biomedcentral.com/articles/10.1186/s13015-015-0042-8). Algorithms in

Molecular Biology 10: 15.



[HS16]

Hordijk, W. and Steel, M.A. (2016). [Autocatalytic sets in polymer networks with variable catalysis distributions](https://www.math.canterbury.ac.nz/~m.steel/Non_UC/files/research/catswim.pdf). J. Math.

Chem., 54(10): 1997-2021.



[GFHKU]

Gatti, R.C., Fath, B., Hordijk, W., Kauffman, S. and Ulanowicz, R. (2018).

[Niche emergence as an autocatalytic process in the evolution of ecosystems](https://pubmed.ncbi.nlm.nih.gov/29864429/),

Journal of Theoretical Biology, 454: 110-117.



[SXH20]

Steel, M., Xavier, J. C., and Huson, D.H. (2020). [The structure of autocatalytic networks, with application to early biochemistry.](https://royalsocietypublishing.org/doi/10.1098/rsif.2020.0488)

J. Royal Society Interface, 17: 20200488.



[XHKSM20]

Xavier, J.C., Hordijk, W., Kauffman, S., Steel M. and Martin, W.F. (2020). [Autocatalytic chemical networks at the origin of metabolism](https://royalsocietypublishing.org/doi/10.1098/rspb.2019.2377). Proc. Roy. Soc. B. 287: 20192377



[H23]

Hordijk, W. (2023). [A concise and formal definition of RAF sets and the RAF algorithm](https://arxiv.org/abs/2303.01809), arXiv:2303.01809.



[HXS24]

Huson, D. H., Xavier, J. C., & Steel, M. A. (2024). _Self‑generating autocatalytic networks: structural results, algorithms, and their relevance to early biochemistry_. *Journal of the Royal Society Interface*, **21**(214), Article 20230732. [📄 PDF](https://royalsocietypublishing.org/doi/pdf/10.1098/rsif.2023.0732) • [🔗 DOI](https://doi.org/10.1098/rsif.2023.0732)



[HXS24b]

Huson, D. H., Xavier, J. C., & Steel, M. A. (2024). _CatReNet: interactive analysis of (auto‑) catalytic reaction networks_. *Bioinformatics*, **40**(8), btae515. [📄 PDF](https://academic.oup.com/bioinformatics/article-pdf/40/8/btae515/58967196/btae515.pdf) • [🔗 DOI](https://doi.org/10.1093/bioinformatics/btae515)