Revisiting Algorithms, Tools, and Applications for Sequence and Phylogenetic Analyses in the NGS-Based Omics Era

Integrating high-throughput sequencing with phylogenetic analysis now spans everything from single genes to long-read pangenomes and metagenomes, yet practitioners still face fragmented, tool-centric guidance. This review revisits algorithms, tools, and workflows for sequence and phylogenetic analysis in the NGS-based omics era, with a focus on comparative performance and scenario-driven decision-making. We first organise classical approaches to tree reconstruction - distance methods, maximum parsimony, maximum likelihood, and Bayesian inference - around core criteria of consistency, efficiency, robustness, and computational cost. We then examine multiple sequence alignment strategies, contrasting progressive, consistency-based, and structure-aware algorithms (such as MAFFT variants and T-Coffee family tools) with segment-based and incremental approaches (for example DIALIGN, anchored domains, and local updates) and alignment-free representations based on k-mers, absent words, and related statistics. For inference, we compare heuristic engines optimised for ultra-large alignments (FastTree, VeryFastTree, online tree optimisation) with full ML frameworks (IQ-TREE, RAxML-NG) and Bayesian platforms for time-scaled phylogenies and phylodynamics (MrBayes, BEAST family). We explicitly discuss trade-offs in accuracy, memory, scalability, and uncertainty support, and show how GPU-enabled implementations change the feasible design space. Beyond these core components, we address current trends that strongly influence method choice: long-read assemblies and pangenomes; data quality issues, contamination, recombination, and horizontal gene transfer; phylogenetic placement and alignment-free screening in metagenomics; and real-time pathogen surveillance using Nextstrain-style workflows. A dedicated section covers workflow management and containerisation (Snakemake, Nextflow, Docker/Singularity) together with benchmarking datasets and FAIR reporting, positioning reproducible pipelines as a first-class requirement rather than an afterthought. To make the review directly actionable, we provide a methodological checklist, a decision framework figure mapping input data to recommended strategies, and a large comparative table summarising algorithmic principles, best use cases, strengths, limitations, scalability, uncertainty support, and reproducibility notes for widely used tools. Applications in infectious disease genomics, oncology, and microbiome research illustrate how these choices translate into biological and clinical insight in practice.