Biology subjects

Rodrigo, A. G.

Publications and source records attributed to Rodrigo, A. G..

10 recordsLinked to original sources

bioRxiv2024

NJGPT: A Large Language Model-Driven, User-Friendly Solution for Phylogenetic Tree Construction

MotivationPhylogenetic reconstruction plays an integral part in much of the research in evolutionary biology. Currently, a newly minted phylogeneticist must choose amongst a relatively large array of phylogenetic software, each often with its own analytical routines, inputs and outputs. Our overarching aim is to construct a user-friendly pipeline with the most recent generative AI tool, ChatGPT-4 (at the time of writing) released by OpenAI, that is able to understand queries written in natural language, to build a phylogenetic tree using sequence data. By doing this, we demonstrate how generative AI may be used in phylogenetics, as a proof-of-concept. We also demonstrate the steps needed, presently, to ensure that ChatGPT can build phylogenetic trees accurately. ResultsWe present NJGPT, a phylogenetic tool built using ChatGPT, a Large Language Model (LLM) Generative Pre-trained Transformer (GPT),which employs the Neighbor-Joining method to construct phylogenetic trees. NJGPT simplifies phylogenetic tree construction by allowing users to generate and visualize trees using natural language queries. It supports multiple sequence file formats, matrix calculation models, and gap-deletion methods. To evaluate the performance of NJGPT, we compared output and runtimes with the widely-used phylogenetic software, MEGA. Our results show that NJGPT produces identical trees over a range of sequence lengths and simple models of evolution. However, NJGPT faces visualization issues with datasets over 50 taxa and operational failures with larger datasets due to token limits. NJGPT runtimes were also substantially slower than MEGA; however, NJGPTs user-friendly interface makes it ideal for beginners. AvailabilityThis plugin is available for free at https://chatgpt.com/g/g-1OzP3Qviw-njgpt, the source code is available on GitHub (https://github.com/ZWan622/NJGPT1.0.git) and is implemented using Python Contactzwan622@aucklanduni.ac.nz Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2024

A digital DNA system reveals the superiority of unidirectional inheritance over 'Lamarckian' inheritance

In biology, changes to a DNA sequence can impact protein sequence but changes to protein sequence (phenotype) do not flow back into DNA (genotype). A system with bidirectional information flow (i.e. both translation and reverse translation) remains a theoretical possibility for an independent origin of life or an artificial biosystem, but the recent development of digital data storage in DNA does just this: changes made to a digital file can be written back into DNA, meaning changes to phenotype can be written back to genotype. To explore the evolutionary properties of such a system, we created an artificial system where synthetic DNA serves as genotype and music as phenotype. Audio can be output from a DNA sequence, then recorded and written to DNA as codons, enabling bidirectional information flow (DNA[->]music and music[->]DNA). Our results show that the mutation rate in a bidirectional system is much higher than for unidirectional information flow, and that, under reverse translation there is no mechanism for preservation of codon choice across generations. This has the effect of eliminating the impact of spontaneous synonymous mutations, a key the benefit of a redundant genetic code. As a result, non-synonymous mutations are the only DNA-level changes that are transmitted across generations, and, as non-synonymous mutation can emerge at both genotypic and phenotypic levels, these occur at a two-fold higher frequency than in a unidirectional system. Our system holds some practical insight. First, for DNA read/write systems, it may be wise to avoid designing systems with de novo reverse translation because the opportunities for mutation are higher; tracking genotype information from the preceding generation to guide this process may reduce error. Second, our system helps clarify how a Lamarckian biological system might operate. We conclude that, were a Lamarckian system of inheritance a feature of early genetic systems, it would likely have been short lived as the high frequency of mutation would risk driving the system to extinction. A system based on unidirectional information flow thus appears superior as there are fewer opportunities for mutational error.

evolutionary biology↗

bioRxiv2024

Joint consideration of selection and microbial generation count provides unique insights into evolutionary and ecological dynamics of holobionts

The relationship between, and joint selection on, a host and its microbes - the holobiont - can impact evolutionary and ecological outcomes of the host and its microbial community. Here we present a novel agent-based modelling framework for understanding the ecological dynamics of hosts and their microbiomes. Our model explicitly incorporates numerous microbial generations per host generation allowing for selection on both host and microbe. We apply our model to explore community fitness and diversity in the face of rapid environmental change. We demonstrate that multiple microbial generations can buffer changes experienced across host lifetimes by smoothing environmental transitions. Our simulations reveal that microbial fitness and host fitness may be at odds with each other when considering the impact of vertical inheritance of microbial communities from a host to its offspring - where high values favour microbial fitness, while low values favour host fitness - these tradeoffs are minimized when microbial generation count per host generation is high. We suggest that these results arise from cross-generational priority effects which maintain diversity within the community and can subsequently be acted upon by selection. Our model is readily extensible into new areas of holobiont research and provides novel insights into holobiont evolution under variable environmental conditions.

microbiology↗

bioRxiv2024

Models of microbiome evolution incorporating environmental microbial selection

The environment plays a crucial role in shaping the microbial composition of hosts. Different environments harbor distinct microbial communities, leading to variation in the availability of different microbes to potential hosts. Over the past few decades, human activities have significantly altered the environment, resulting in a decline in microbial diversity. Currently, the impact of this environmentally-induced reduction in microbial diversity on environmental microbial communities, host-microbiome interactions, and their evolutionary implications remains unclear. In this paper, we simulated how environmental microbial selection can lead to decreasing microbial diversity, using a previously-developed selection model. Our results indicate that microbial communities are influenced by both environmental conditions and host preferences. The diversity of the microbiome carried by hosts is impacted by the environment, with stronger environmental microbial selection reducing microbial diversity in the hosts, while at the same time, facilitating the emergence of mutually beneficial symbioses between microorganisms and hosts. Environmental microbial selection (EMS) leads to the survival of highly fit microorganisms in the environment, while the fitnesses of hosts and microbes within hosts remain relatively unaffected by environmental microbial selection.

evolutionary biology↗

bioRxiv2024

Models of microbiome evolution incorporating host resource provisioning

Multicellular hosts and their associated microbial partners (i.e., microbiomes) often interact in mutually beneficial ways. Consequently, hosts may choose to allocate resources to regulate and recruit appropriate microbes. In doing so, hosts may incur an energetic cost, and in turn, these costs can affect host fitness. It remains unclear how hosts have evolved to balance the costs of expending resources to manage their microbiomes against the benefits that might accrue by doing so. We extended a previously-developed agent-based computational model of hostmicrobiome evolution by incorporating a resource provisioning process, whereby hosts provide resources to support beneficial microbes and suppress harmful microbes, with attendant fitness costs. Our results indicate that the ways and sources from which a host acquires microbes are crucial factors influencing the hosts willingness to provide resources, to regulate microbiome composition. The intensity of resource provisioning will depend, in part, on how much of their microbiome hosts contribute to, and obtain from, their environment: when hosts that engage in resource provisioning contribute a high percentage of their microbiome to the environment, then there is less evolutionary imperative for other hosts to also provide resources. Since resource provisioning incurs a fitness cost to the host, over evolutionary time, resource provisioning will not be favored. Additionally, if selection at the microbial level is not sufficiently strong, and the host obtains a high proportion of microbes from the environment, then the higher the proportion of beneficial microbes in the environment, the less the host is willing to provide resources.

microbiology↗

bioRxiv2023

Self-generating autocatalytic networks: structural results, algorithms, and their relevance to evolutionary processes

The concept of an autocatalytic network of reactions that can form and persist, starting from just an available food source, has been formalised by the notion of a Reflexively-Autocatalytic and Food generated (RAF) set. The theory and algorithmic results concerning RAFs have been applied to a range of settings, from metabolic questions arising at the origin of life, to ecological networks, and cognitive models in cultural evolution. In this paper, we present new structural and algorithmic results concerning RAF sets, by studying more complex modes of catalysis that allow certain reactions to require multiple catalysts (or to not require catalysis at all), and discuss the differing ways catalysis has been viewed in the literature. We then focus on the structure and analysis of minimal RAFs, and derive structural results and polynomial-time algorithms, with applications to metabolic network data described briefly.

evolutionary biology↗

bioRxiv2022

MAST: Phylogenetic Inference with Mixtures Across Sites and Trees

Hundreds or thousands of loci are now routinely used in modern phylogenomic studies. Concatenation approaches to tree inference assume that there is a single topology for the entire dataset, but different loci may have different evolutionary histories due to incomplete lineage sorting, introgression, and/or horizontal gene transfer; even single loci may not be treelike due to recombination. To overcome this shortcoming, we introduce the mixture across sites and trees (MAST) model, which uses a mixture of bifurcating trees to represent multiple histories in a single concatenated alignment. The MAST model allows each tree to have its own topology, branch lengths, substitution model, nucleotide or amino acid frequencies, and model of rate heterogeneity across sites. We implemented the MAST model in a maximum-likelihood framework in the popular phylogenetic software, IQ-TREE. Simulations show that we can accurately recover the true model parameters, including branch lengths and tree weights (i.e. frequencies) for a given set of tree topologies. We also show that we can use standard statistical inference approaches to reject a single-tree model when data are simulated under multiple trees (and vice versa). We applied the MAST model to multiple primate datasets and found that it can recover the signal of incomplete lineage sorting in the Great Apes, as well as the asymmetry in minor trees caused by introgression among several macaque species. When applied to a dataset of four Platyrrhine species for which standard concatenated maximum likelihood and gene tree approaches disagree, we find that MAST gives the highest weight to the tree favored by gene tree approaches. These results suggest that the MAST model is able to analyse a concatenated alignment using maximum likelihood, while avoiding some of the biases that come with assuming there is only a single tree. The MAST model can therefore offer unique biological insights when applied to datasets with multiple evolutionary histories. We discuss how it can be extended in the future.

bioinformatics↗

bioRxiv2021

An assembly-free method of phylogeny reconstruction using short-read sequences from pooled samples without barcodes

A current strategy for obtaining haplotype information from several individuals involves short-read sequencing of pooled amplicons, where fragments from each individual is identified by a unique DNA barcode. In this paper, we report a new method to recover the phylogeny of haplotypes from short-read sequences obtained using pooled amplicons from a mixture of individuals, without barcoding. The method, AFPhyloMix, accepts an alignment of the mixture of reads against a reference sequence, obtains the single-nucleotide-polymorphisms (SNP) patterns along the alignment, and constructs the phylogenetic tree according to the SNP patterns. AFPhyloMix adopts a Bayesian model of inference to estimates the phylogeny of the haplotypes and their relative frequencies, given that the number of haplotypes is known. In our simulations, AFPhyloMix achieved at least 80% accuracy at recovering the phylogenies and frequencies of the constituent haplotypes, for mixtures with up to 15 haplotypes. AFPhyloMix also worked well on a real data set of kangaroo mitochondrial DNA sequences.

bioinformatics↗

bioRxiv2018

Effective Machine-Learning Assembly For Next-Generation Sequencing With Very Low Coverage

In short-read DNA sequencing experiments, the read coverage is a key parameter to successfully assemble the reads and reconstruct the sequence of the input DNA. When coverage is very low, the original sequence reconstruction from the reads can be difficult because of the occurrence of uncovered gaps. Reference guided assembly can then improve these assemblies. However, when the available reference is phylogenetically distant from the sequencing reads, the mapping rate of the reads can be extremely low. Some recent improvements in read mapping approaches aim at modifying the reference according to the reads dynamically. Such approaches can significantly improve the alignment rate of the reads onto distant references but the processing of insertions and deletions remains challenging. Here, we introduce a dynamic programming algorithm to update the reference sequence according to previously aligned reads. Substitutions, insertions and deletions are performed in the reference sequence dynamically. We evaluate this approach to assemble a western-grey kangaroo mitochondrial amplicon. Our results show that more reads can be aligned and that this method produces assemblies of length comparable to the truth while limiting error rate when classic approaches fail to recover the correct length. Our method allows us to assemble the first full mitochondrial genome for the western-grey kangaroo. Finally, we discuss how the core algorithm of this method could be improved and combined with other approaches to analyse larger genomic sequences.

bioinformatics↗

bioRxiv2018

HaploJuice : Accurate haplotype assembly from a pool of sequences with known relative concentrations

Pooling techniques, where multiple sub-samples are mixed in a single sample, are widely used to take full advantage of high-throughput DNA sequencing. Recently, Ranjard et al. [1] proposed a pooling strategy without the use of barcodes. Three sub-samples were mixed in different known proportions (i.e. 62.5%, 25% and 12.5%), and a method was developed to use these proportions to reconstruct the three haplotypes effectively. HaploJuice provides an alternative haplotype reconstruction algorithm for Ranjard et al.s pooling strategy. HaploJuice significantly increases the accuracy by first identifying the empirical proportions of the three mixed sub-samples and then assembling the haplotypes using a dynamic programming approach. HaploJuice was evaluated against five different assembly algorithms, Hmmfreq [1], ShoRAH [2], SAVAGE [3], PredictHaplo [4] and QuRe [5]. Using simulated and real data sets, HaploJuice reconstructed the true sequences with the highest coverage and the lowest error rate. HaploJuice achieves high accuracy in haplotype reconstruction, making Ranjard et al.s pooling strategy more efficient, feasible, and applicable, with the benefit of reducing the sequencing cost.

bioinformatics↗