betaFeedback

WorksMathematics

A Central Limit Theorem for the Driver Site Frequency Spectrum

Abstract

We consider a supercritical two-type continuous-time linear birth-death process with mutation and selection, in which wild-type individuals give rise to mutant offspring with a larger net-growth rate. We previously had obtained first-order law of large number results for the driver site frequency spectrum. We return to obtain a second-order central limit theorem, in which we show stable convergence of a scaling of the deviation of the SFS from its first-order limit. We provide a central limit theorem at both large detection times and sizes. Then, we apply these results to obtain confidence intervals for a class of consistent estimators for the fitness advance.

Introduction

Cells can accumulate mutations in a sequential manner, through which it becomes cancerous [1, 2]. One way to summarize the data concerning the frequency of these mutations is through the site frequency spectrum (denoted (Sj(t))j=1∞(S_j(t))_{j=1}^{\infty}). (Sj(t))j=1∞(S_j(t))_{j=1}^{\infty} is a random vector with the jjth coordinate denoting the number of mutations present in jj cells at time tt. These mutations can either be deleterious, neutral, or beneficial. Here, we consider the setting of driver (i.e. beneficial) mutations which is relevant in studying the development of cancer, as in our previous work [3].

Specifically, the model we consider is a two-type continuous-time linear birth-death process, wherein type-0 cells can mutate to type-1 cells with a fitness increase in its birth rate. Under this model, we had previously obtained a law of large numbers concerning Sj(t)S_j(t) and an associated estimator for the fitness increase of the type-1 clone over the type-1 clone [3]. Here, we provide second-order asymptotics, or a central limit theorem, for the SFS and use it to construct confidence intervals for our estimator.

Prior works have shown first-order-type results for the SFS under deleterious [4], neutral [5, 6, 7, 4, 8, 9], or selective/rescue [4, 10, 3] evolution. Second-order results are somewhat harder to come-by. By exploiting recently-developed coalescent tools for a sample taken from a Galton-Watson process [11, 12], Schweinsberg and Shuai [9] have derived a central limit theorem for a sample of the SFS in the neutral mutation setting. They use a backwards-in-time approach exploiting the coalescent structure of a sample taken from a Galton-Watson tree [11], whereas in this study we use a different forward-in-time approach to obtain a central limit theorem.

Outline

In § 3.1, we prove one of our main results concerning the convergence of the scaled deviations of Sj(t)S_j(t). We also examine the large time deviations when j=j(t)j=j(t) depends on tt. In § 3.2, we provide a law of iterated logarithm result for this Gaussian process. We extend the CLT when detecting at detection sizes in § 3.3. Finally in Section 3.4, we provide an application of our results to constructing confidence intervals for an estimator for the fitness advance constructed in our previous work [3]. Most proofs are deferred to § 5.

Model Description

Consider a continuous-time Markov branching process with two cell types, denoted type-0 and type-1. Each type-0 cell independently divides at exponential rate aa and dies at rate dd, and the total type-0 population count at time tt is denoted Z0(t)Z_0(t). Each type-1 cell divides at rate a+ba+b and die at rate dd, where b>0b>0 reflects a selective advantage. The total type-1 population count at time tt is denoted Z1(t)Z_1(t). Type-1 cells are created via driver mutation acquisition. The total population is denoted Z(t)=Z0(t)+Z1(t)Z(t)=Z_0(t)+Z_1(t)

Type-0 cells produce type-1 cells through two mechanisms, (i) replication-dependent: at each type-0 division, precisely one offspring is type-1 with probability vv and type-0 with probability 1−v1-v1, and (ii) replication-independent: each type-0 cell mutates to type-1 at rate μ\mu, independently of division [13]. Thus, the effective birth rate of type-0 cells is a(1−v)a(1-v) and the net growth rate is λ0≡a(1−v)−d−μ\lambda_0 \equiv a(1-v)-d-\mu. The net growth rate of type-1 cells is λ1≡a+b−d\lambda_1 \equiv a+b-d. Throughout, we assume supercriticality of the type-0 population, namely λ0>0\lambda_0>0. The non-extinction event of the type-0 population is denoted Ω0∞={Z0(t)>0 ∀t}\Omega^\infty_0=\{Z_0(t)>0\ \forall t\}. It is well known that P(Ω0∞):=q0=1−p0\mathbb{P}(\Omega^\infty_0):=q_0=1-p_0, where p0=(d+μ)/(a(1−v))p_0=(d+\mu)/(a(1-v)). The probability of extinction of the type-1 population, conditional on starting with a single type-1 cell, is p1=d/(a+b)p_1=d/(a+b). We also let q1=1−p1q_1=1-p_1 denote its probability of non-extinction.

We denote the transition probabilities of a single type-1 clone by Pj(t)=P(Z1(t)=j∣Z1(0)=i)P_j(t)=\mathbb{P}(Z_1(t)=j\mid Z_1(0)=i). Let Sj(t)S_j(t) denote the number of type-1 driver clones of size jj at time tt; we refer to Sj(t)S_j(t) as the ‘driver SFS.’ Specifically, letting Z1(i)(s)Z_1^{(i)}(s) denote the size of the iith type-1 clone ss units after it is initiated:

Sj(t)=∑i:Ti≤t1Z1(i)(t−Ti)=jS_j(t)=\sum_{i:T_i\le t}\mathbf{1}_{Z_1^{(i)}(t-T_i)=j}

We also denote Mt=#{i:Ti≤t}M_t=\#\{i:T_i\le t\} as the number of mutation events from type-0 to type-1 that have occurred in the interval [0,t][0,t]. We explore the SFS at both large detection times as well as large detection times τn=inf⁡{t≥0:Z(t)≥n}\tau_n=\inf\{t\ge0:Z(t)\ge n\}.

It is well-known that lim⁡t→∞e−λ0tZ0(t)\lim_{t\to\infty}e^{-\lambda_0t}Z_0(t) converges almost surely and in L2L^2 to a limiting random variable YY. The limiting random variable YY is 00 with probability p0p_0, and on the event of survival it is exponential with rate q0q_0 (for example see Theorem 1 in [14]). Also, W=lim⁡t→∞e−λ1tZ(t)=lim⁡t→∞e−λ1tZ1(t)W=\lim_{t\to\infty}e^{-\lambda_1t}Z(t)=\lim_{t\to\infty}e^{-\lambda_1t}Z_1(t) almost surely (see Lemma 2).

We will be showing stable convergence with respect to certain σ\sigma-algebras, and so we make reference to certain important σ\sigma-filtrations

  • Let Ft≤i\mathcal{F}^{\le i}_t for i∈{0,1}i\in\{0,1\} denote the smallest σ\sigma-algebra generated either (if i=0i=0) by the type-0 process in the time interval [0,t][0,t] or (if i=1i=1) by the type-0 and type-1 process in the time interval [0,t][0,t]. The i=0i=0 case is most often used, so that we also write Ft:=Ft≤0\mathcal{F}_t:=\mathcal{F}^{\le0}_t. We also let F∞≤i=σ(⋃t≥0Ft≤i)\mathcal{F}^{\le i}_\infty=\sigma\left(\bigcup_{t\ge0}\mathcal{F}^{\le i}_t\right). Once again, for shorthand we write F∞:=F∞≤0\mathcal{F}_\infty:=\mathcal{F}^{\le0}_\infty.

  • Let Bt≤i\mathcal{B}^{\le i}_t for i∈{0,1}i\in\{0,1\} denote the smallest σ\sigma-algebra containing Ft≤i\mathcal{F}^{\le i}_t as well as the timings of the type-1 mutation times that occur within the interval [0,t][0,t]. We also write Bt:=Bt≤0\mathcal{B}_t:=\mathcal{B}^{\le0}_t. We also let B∞≤i=σ(⋃t≥0Bt≤i)\mathcal{B}^{\le i}_\infty=\sigma\left(\bigcup_{t\ge0}\mathcal{B}^{\le i}_t\right). Once again, for shorthand we write B∞:=B∞≤0\mathcal{B}_\infty:=\mathcal{B}^{\le0}_\infty.

  • Let It\mathcal{I}_t be the smallest σ\sigma-algebra containing Bt≤1\mathcal{B}^{\le1}_t as well as the identities of the clones to which each type-1 cell belongs to. Finally, define I∞=σ(⋃t≥0It)\mathcal{I}_\infty=\sigma\left(\bigcup_{t\ge0}\mathcal{I}_t\right).

Clearly, Ft≤i⊆Bt≤i⊆It\mathcal{F}_t^{\le i} \subseteq\mathcal{B}_t^{\le i} \subseteq\mathcal{I}_t. (Z0(t),Z1(t))(Z_0(t), Z_1(t)) is adapted to Ft≤1\mathcal{F}_t^{\le1}. (Z0(t),Z1B(t))(Z_0(t), Z_1^{B}(t)) and Sj(t)S_j(t) are adapted to It\mathcal{I}_t. For a sub-σ\sigma-algebra G\mathcal{G} of our ambient probability space, G\mathcal{G}-stable convergence of {Qt}t\{Q_t\}_t to Q∞Q_\infty means that for all bounded G\mathcal{G}-measurable Q′Q' and continuous, bounded ff:

E[Q′f(Qt)]→E[Q′f(Q∞)].\mathbb{E}[Q'f(Q_t)] \to\mathbb{E}[Q'f(Q_\infty)].

Letting ℓ∈R∪{±∞}\ell\in\mathbb{R} \cup\{\pm\infty\} and g,h:R→R≥0g,h:\mathbb{R}\to\mathbb{R}_{\ge0}, we also use the following asymptotic notation:

g∼t→ℓh⟺lim⁡t→ℓg(t)h(t)=1g \mathop{\sim}_{t\to\ell} h \Longleftrightarrow\lim_{t\to\ell}\frac{g(t)}{h(t)}=1
g≲t→ℓh⟺lim sup⁡t→ℓg(t)h(t)<∞g \mathop{\lesssim}_{t\to\ell} h \Longleftrightarrow\limsup_{t\to\ell}\frac{g(t)}{h(t)}<\infty
g≳t→ℓh⟺h≲t→ℓgg \mathop{\gtrsim}_{t\to\ell} h \Longleftrightarrow h \mathop{\lesssim}_{t\to\ell} g
g≪t→ℓh⟺lim⁡t→ℓg(t)h(t)=0g \mathop{\ll}_{t\to\ell} h \Longleftrightarrow\lim_{t\to\ell}\frac{g(t)}{h(t)}=0
g≫t→ℓh⟺h≪t→ℓgg \mathop{\gg}_{t\to\ell} h \Longleftrightarrow h \mathop{\ll}_{t\to\ell} g
g≍t→ℓh⟺g≲t→ℓh and g≳t→ℓhg \mathop{\asymp}_{t\to\ell} h \Longleftrightarrow g \mathop{\lesssim}_{t\to\ell} h \text{ and } g \mathop{\gtrsim}_{t\to\ell} h

Additionally, the Landau-O notation is used to specify in-line error terms. That is, letting g:R→Rg:\mathbb{R}\to\mathbb{R} and h:R→R≥0h:\mathbb{R}\to\mathbb{R}_{\ge0} be real-functions, we say:

g=Ot→ℓ(h)⟺∣g∣≲t→ℓhg = O_{t\to\ell}(h) \Longleftrightarrow|g| \mathop{\lesssim}_{t\to\ell} h
g=ot→ℓ(h)⟺∣g∣≪t→ℓhg = o_{t\to\ell}(h) \Longleftrightarrow|g| \mathop{\ll}_{t\to\ell} h

Results

Central Limit Theorem of the SFS

Our first result is a weak convergence result for the net of centered and scaled SFS vectors in the large time limit. This result is strengthened to establish I∞\mathcal{I}_\infty-stable convergence in order to later prove a weak convergence result in the large detection size limit.

References

  1. [1]Jason H. Bielas et al. “Human cancers express a mutator phenotype”. In: Proceedings of the National Academy of Sciences 103.48 (2006), pp. 18238–18242. DOI: 10 . 1073 / pnas . 0607057103. eprint: https://www.pnas.org/doi/pdf/10.1073/pnas.0607057103. URL: https://www.pnas.org/doi/10.1073/pnas.0607057103.DOI
  2. [2]Douglas Hanahan. “Hallmarks of Cancer: New Dimensions”. In: Cancer Discovery 12.1 (Jan. 2022), pp. 31–46. ISSN: 2159-8274. DOI: 10 . 1158 / 2159–8290 . CD–21–1059. eprint: https://aacrjournals.org/cancerdiscovery/article-pdf/12/1/31/3052722/31.pdf. URL: https://doi.org/10.1158/2159–8290.CD–21–1059.
  3. [3]Armaan Ahmed et al. The Site Frequency Spectrum in an Exponentially-Growing Population with Selection. 2026. arXiv: 2607 . 16479 [math.PR]. URL: https://arxiv.org/abs/2607.16479.
  4. [4]David Cheek and Tibor Antal. “MUTATION FREQUENCIES IN A BIRTH–DEATH BRANCH-ING PROCESS”. In: The Annals of Applied Probability 28.6 (2018), pp. 3922–3947. ISSN: 10505164, 21688737. URL: https://www.jstor.org/stable/26542506 (visited on 10/15/2025).arxiv.org/abs/1710.09783
  5. [5]Rick Durrett. “Population genetics of neutral mutations in exponentially growing cancer cell populations”. In: The Annals of Applied Probability 23.1 (2013), pp. 230–250. DOI: 10.1214/11–AAP824. URL: https://doi.org/10.1214/11–AAP824.
  6. [6]Ivana Bozic, Jeffrey M. Gerold, and Martin A. Nowak. “Quantifying Clonal and Subclonal Passenger Mutations in Cancer Evolution”. In: PLOS Computational Biology 12.2 (Feb. 2016), pp. 1–19. DOI: 10.1371/journal.pcbi.1004731. URL: https://doi.org/10.1371/journal.pcbi.1004731.
  7. [7]Marc J. Williams et al. “Identification of neutral tumor evolution across cancer types”. In: Nature Genetics 48.3 (Mar. 2016), pp. 238–244. ISSN: 1546-1718. DOI: 10.1038/ng.3489. URL: https://doi.org/10.1038/ng.3489.
  8. [8]Einar Bjarki Gunnarsson, Kevin Leder, and Xuanming Zhang. “Limit theorems for the site frequency spectrum of neutral mutations in an exponentially growing population”. In: Stochastic Processes and their Applications 182 (2025), p. 104565. ISSN: 0304-4149. DOI: https://doi.org/10.1016/j.spa.2025.104565. URL: https://www.sciencedirect.com/science/article/pii/S0304414925000043.
  9. [9]Jason Schweinsberg and Yubo Shuai. “Asymptotics for the site frequency spectrum associated with the genealogy of a birth and death process”. In: The Annals of Applied Probability 35.1 (2025), pp. 99–141. DOI: 10.1214/24–AAP2110. URL: https://doi.org/10.1214/24–AAP2110.
  10. [10]Céline Bonnet and Hélène Leman. “Site frequency spectrum of a rescued population under rare resistant mutations”. In: Stochastic Processes and their Applications 176 (2024), p. 104421. ISSN: 0304-4149. DOI: https://doi.org/10.1016/j.spa.2024.104421. URL: https://www.sciencedirect.com/science/article/pii/S0304414924001273.
  11. [11]Amaury Lambert. “The coalescent of a sample from a binary branching process”. In: Theoretical Population Biology 122 (2018), Paul Joyce, pp. 30–35. ISSN: 0040-5809. DOI: https://doi.org/10.1016/j.tpb.2018.04.005. URL: https://www.sciencedirect.com/science/article/pii/S0040580918300753.
  12. [12]Simon C. Harris, Samuel G. G. Johnston, and Matthew I. Roberts. “THE COALESCENT STRUCTURE OF CONTINUOUS-TIME GALTON–WATSON TREES”. In: The Annals of Applied Probability 30.3 (2020), pp. 1368–1414. ISSN: 10505164, 21688737. URL: https://www.jstor.org/stable/26965977 (visited on 12/23/2025).arxiv.org/abs/1703.00299
  13. [13]Federico Abascal et al. “Somatic mutation landscapes at single-molecule resolution”. In: Nature 593.7859 (May 2021), pp. 405–410. ISSN: 1476-4687. DOI: 10.1038/s41586–021–03477–4. URL: https://doi.org/10.1038/s41586–021–03477–4.
  14. [14]Richard Durrett. Branching Process Models of Cancer. en. Cham: Springer International Publishing, 2015. ISBN: 978-3-319-16064-1 978-3-319-16065-8. DOI: 10.1007/978-3-319-16065-8. URL: https://link.springer.com/10.1007/978-3-319-16065-8 (visited on 10/05/2025).DOI
  15. [15]Jean Jacod and Albert N. Shiryaev. Limit Theorems for Stochastic Processes. en. Ed. by A. ChenCiner et al. Vol. 288. Grundlehren der mathematischen Wissenschaften. Berlin, Heidelberg: Springer, 2003. ISBN: 978-3-642-07876-7 978-3-662-05265-5. DOI: 10.1007/978-3-662-05265-5. URL: http://link.springer.com/10.1007/978-3-662-05265-5 (visited on 05/11/2026).DOI
  16. [16]M. R. Leadbetter, Georg Lindgren, and Holger Rootzén. Extremes and Related Properties of Random Sequences and Processes. Springer Series in Statistics. New York, NY: Springer, 1983. ISBN: 978-1-4612-5451-5 978-1-4612-5449-2. DOI: 10.1007/978-1-4612-5449-2. URL: http://link.springer.com/10.1007/978-1-4612-5449-2 (visited on 07/14/2026).DOI

Paper details

vlnc:W_z7uk3vx6
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents