<i>Blang</i>: Bayesian Declarative Modeling of General Data Structures and Inference via Algorithms Based on Distribution Continua

Bouchard‐Côté, Alexandre; Chern, Kevin; Čubranić, Davor; Hosseini, Sahand; Hume, J. N. P.; Lepur, Matteo; Ouyang, Zihui; Sgarbi, Giorgio

doi:10.18637/jss.v103.i11

Cited by 6 publications

(5 citation statements)

References 47 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For instance, one might be interested in eco-evolutionary models where the lineages interact so that some lineage-specific parameters are dependent on the state of all other lineages in the community. As pointed out in [12], such problems can be addressed in a PPL setting by proposing an initial state from a simpler model, and then weighting the simulation by the probability of the full model, as we exemplified in the host repertoire and tree inference programs.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

TreePPL: A Universal Probabilistic Programming Language for Phylogenetics

Senderov,

Kudlicka,

Lundén

et al. 2023

Preprint

View full text Add to dashboard Cite

We present TreePPL, a language for probabilistic modeling and inference in statistical phylogenetics. Specifically, TreePPL is a domain-specific universal probabilistic programming language (PPL), particularly designed for describing phylogenetic models. The core idea is to express the model as a computer program, which estimates the posterior probability distribution of interest when executed sufficiently many times. The program uses two special probabilistic constructs:assumestatements, which describe latent random variables in the model, andobservestatements, which condition random variables in the model on observed data. Theassumeandobservestatements make it possible for generic inference algorithms, such as sequential Monte Carlo and Markov chain Monte Carlo algorithms, to identify checkpoints that enable them to generate and manipulate simulations from the posterior probability distribution. This means that a user can focus on describing the model, and leave the estimation of the posterior probability distribution to TreePPL’s inference machinery. The TreePPL modeling language is inspired by R, Python, and the functional programming language OCaml. The model script can be conveniently run from a Python environment (an R environment is work in progress), which can be used for pre-processing, feeding the model with the observed data, controlling and running the inference, and receiving and post-processing the output data. The inference machinery is generated by a compiler framework developed specifically for supporting domain-specific modeling and inference, the Miking CorePPL framework. It currently supports a range of inference strategies, including several recent innovations that are important for efficient inference on phylogenetic models. It also supports the implementation of novel inference strategies for models described using TreePPL or other domain-specific modeling languages. We briefly describe the TreePPL modeling language and the Python environment, and give some examples of modeling and inference with TreePPL. The examples illustrate how TreePPL can be used to address a range of common problem types considered in statistical phylogenetics, from diversification and co-speciation analysis to tree inference. Although much progress has been made in recent years, developing efficient algorithms for automatic PPL-based inference is still a very active field. A few major challenges remain to be addressed before the entire phylogenetic model space is adequately covered by efficient automatic inference techniques, but several of them are being addressed in ongoing work on TreePPL. We end the paper by discussing how probabilistic programming can support the use of machine learning in designing and fine-tuning inference strategies and in extending incomplete model descriptions in phylogenetics.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Very recently, we have seen additional phylogenetic modeling languages introduced [9, 10]. There has also been a growing interest in leveraging generic statistical modeling and inference frameworks for the analysis of phylogenetic models [11, 12].…”

Section: Introductionmentioning

confidence: 99%

TreePPL: A Universal Probabilistic Programming Language for Phylogenetics

Senderov,

Kudlicka,

Lundén

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The posterior distribution is summarized using a Bayes estimator described in Section 9.4.5. The model is implemented in the Blang probabilistic programming language [35].…”

Section: Methodsmentioning

confidence: 99%

Cancer phylogenetic tree inference at scale from 1000s of single cell genomes

Dorri

Salehi²,

Chern³

et al. 2020

Preprint

View full text Add to dashboard Cite

A new generation of scalable single cell whole genome sequencing (scWGS) methods [Zahn et al., 2017, Laks et al., 2019, allows unprecedented high resolution measurement of the evolutionary dynamics of cancer cells populations. Phylogenetic reconstruction is central to identifying sub-populations and distinguishing mutational processes. The ability to sequence tens of thousands of single genomes at high resolution per experiment [Laks et al., 2019] is challenging the assumptions and scalability of existing phylogenetic tree building methods and calls for tailored phylogenetic models and scalable inference algorithms. We propose a phylogenetic model and associated Bayesian inference procedure which exploits the specifics of scWGS data. A first highlight of our approach is a novel phylogenetic encoding of copy-number data providing an attractive statistical-computational trade-off by simplifying the site dependencies induced by rearrangements while still forming a sound foundation to phylogenetic inference. A second highlight is an innovative phylogenetic tree exploration move which makes the cost of MCMC iterations bounded by O(|C| + |L|), where |C| is the number of cells and |L| is the number of loci. In contrast, existing off-the-shelf likelihood-based methods incur iteration cost of O(|C| |L|). Moreover, the novel move considers an exponential number of neighbouring trees whereas offthe-shelf moves consider a polynomial size set of neighbours. The third highlight is a novel * Equal contribution 1 mutation calling method that incorporates the copy-number data and the underlying phylogenetic tree to overcome the missing data issue. This framework allows us to realistically consider routine Bayesian phylogenetic inference at the scale of scWGS data.

show abstract

“…The posterior distribution is summarized using a Bayes estimator described in Section 2.4.5. The model is implemented in the Blang probabilistic programming language (Bouchard-Côté et al, 2022).…”

Section: Inference the Posterior Distributionmentioning

confidence: 99%