Steady states for crossing changes in knot diagrams

Harrison Chapman
Colorado State University

hchapman.org/talks/birs_steadystate

The Topology of Nucleic Acids:
Research at the Interface of Low-Dimensional Topology,
Polymer Physics and Molecular Biology

Banff International Research Station
Banff, Alberta
March 24th 2019

Knot Diagrams

Crossings are the primary components of knot diagram models

Crossings as Self-contacts

Crossings may be viewed as self-contacts where enzymes may act to change DNA topology

Strand passage action, e.g. topo-IV

Coherent smoothing action, e.g. XerCD-dif-FtsK complex

Hua, Nguyen, Raghavan, Arsuaga, Vasquez (2007)

Calculated transition probabilities between knot types and steady-state probabilities of DNA molecules under the action of topo-IV.

Issue

Action of topo-IV is easily expressed diagrammatically, but complex to represent with self-avoiding polygons (SAPs) alone

Strategy

Sample SAPs, then transfer to diagrams to study action of topo-IV

Strategy of Hua et al.

For each prime knot type $$K$$, sample diagrams representing $$K$$ from a weighted distribution:

1. Sample SAPs representing $$K$$ uniformly of desired sizes, using BFACF
2. Randomly project to 50–100 diagrams
3. Result is a list of diagrams for $$K$$, weighted by occurrence: sample from this list uniformly

Idea

Crossings in diagrams drawn under weighted distribution likely represent crossings close in original SAPs

Question

How tied are the results of Hua et al. to the projection-weighted distribution on diagrams?

Aim to answer this by applying the same techniques, now using uniform distribution on knotted diagrams

Diagram MCMC

One step of a diagram Markov chain takes as input a knot diagram, performs with some probability a (Reidemeister) transition, and returns the resulting knot diagram

Shadow Markov Chain

Explore all knot shadows by ignoring crossing signs; get all knot diagrams by adding crossing information

Fixed Knot Type Markov Chain

Explore all diagrams of fixed knot type by respecting crossing signs (c.f. BFACF and lattice polygons)

C–Rechnitzer

The Wang-Landau Markov chain on diagrams of knot type $$K$$ has stationary distribution where the probability that an $$n$$-crossing diagram $$D$$ representing $$K$$ is sampled is, $\pi(D) \propto \frac{1}{g_n(K)} \approx \frac{1}{k_n(K)}.$

Knot diagrams for $$K$$ are sampled

• uniformly for any given size, and
• approximately each size is equally likely

Under the (Sampling) Hood

• Maximum diagram size considered: $$70$$ crossings
• Enumeration step ran with flatness parameter $$\Delta = 0.99$$ until update parameter $$f \le 10^{-6}$$
• Enumeration step took approximately 1d to tune all 36 knot types
• Samples drawn every $$1.5 \times 10^{5}$$ attempted transitions
• $$100,000$$ diagrams sampled for each knot type (approx. $$1,500$$ samples per diagram size)
• Data for each knot type gathered in approximately 25m, totalling 15h

Approximate Enumeration: $$6_1$$

$$n$$ Rooted Rooted/$$4n$$ Exact (CCM '16)
6 12 .5 1
7 911.68 32.56 34
8 40 207.17 1 256.47 1 267
9 1 370 185.44 38 060.71 38 199
10 40 659 767.57 1 016 494.19 1 015 996
11 1 103 328 823.05 25 075 655.07 N/A

Exact enumeration data from Cantarella-C-Mastin '16

Strategy for Uniformly Sampled Diagrams

Now we just:

1. For each prime knot type $$K$$, sample diagrams representing $$K$$ from the uniform distribution of a desired size range, using fixed-knot type diagram MCMC

Then, we apply the same analysis as in Hua et al.

Steady States for Topo-IV Action

In either study, after sampling a diagram of type $$K$$:

1. Reduce the diagram fully by $$\mathrm{RI}^{-}$$ and $$\mathrm{RII}^-$$ moves
(except in the case of the unknot)

Steady States for Topo-IV Action

In either study, after sampling a diagram of type $$K$$:

1. Reduce the diagram fully by $$\mathrm{RI}^{-}$$ and $$\mathrm{RII}^-$$ moves
(except in the case of the unknot)
2. Pick a crossing uniformly at random and make a crossing change
3. If the resulting knot type is sufficiently small ($$cr(L) \le 8$$) and prime, record it

After normalizing, result is a vector of transition probabilities for knot type $$K$$

Steady States for Topo-IV Action

Gathering data for each prime knot type $$K$$ of up to 8 crossings yields a square stochastic matrix $$M$$

The eigenvector associated to the largest eigenvalue of $$M$$ is the steady state of the system.

Rescaling the $$x$$-axis?

The expected average crossing number over all projections of a $$M$$-edge Gaussian random polygon is $$O(M \ln M)$$

Can rescale $$x$$-axis by $$e^{W_0(n)}$$ where $$W_0$$ is the main branch of the Lambert $$W$$ function (satisfying $$z = W(ze^z)$$)

Questions and Take-aways

1. Do these data show a key difference between diagrams and space polygons?
2. Uniform diagram sampler allows new tests for null hypotheses in many other previous experiments (e.g. Stolz et al. '17 work on XerCD-dif-FtsK in E. coli)
3. Wang-Landau approximate enumeration provides new enumeration of numbers of diagrams for given knot types