Table 1. Summary of nine approaches included in this study: input required for each approach is either FASTQ (sequencing reads), or VCF (allele frequency table relative to the Wuhan reference). The database used by each approach was divided into two categories: curated mutation profiles, where a selection of alleles present per lineage is done by a functional or phylogenetic criteria, or selected alleles from GISAID genomes, where all or some GISAID genomes are aligned to the reference genome, and alleles are selected based on allele-frequency. For VLQ, this is used in selecting a reference set for lineages when performing pseudoalignment. Database availability refers to how users can access the database, by accessing the tools’ GitHub repository or by executing a command within the tool. For VLQ, this is used in selecting a reference set for lineages when performing pseudoalignment. Database availability refers to how users can access the database, by accessing the tools’ GitHub repository or by executing a command within the tool.

Tool	Package manager	Dependencies	Input	Method to estimate lineage relative abundances	Database type	Database details	Database availability	Reference
Alcov	Pip	Cutadapt, minimap2 Python Packages: Fire, numpy, pandas, scikit-learn, matplotlib, seaborn, pysam	FASTQ	Linear model: ordinary least squared	Curated mutation profiles	Cov-Spectrum	GitHub	[31]
Basic	na	blast+, PRINSEQ, fastp, bwa, picard, samtools, varscan, python3, R	FASTQ	Linear model: constrained linear model	Selected alleles from GISAID genomes	Allele-frequency>=90 %	GitHub	[30]
Freyja	Conda	iVar, samtools, UShER Python packages: cvxpy, numpy, pandas	FASTQ	Constrained (weighted) least absolute deviations	Curated mutation profiles	Marker mutations from UShER global phylogenetic tree	GitHub or; Via Freyja	[10]
Gromstole	na	cutadapt, minimap2, Python, R	FASTQ	Quasibinomial regression model	Curated mutation profiles	Cov-Spectrum	GitHub	[33]
LCS	na	Snakemake, samtools, GATK4, bwa, picard, USHER Python Packages:, biopython, pysam, pyvcf, pandas, cvxpy, ray-core	FASTQ or VCF	Log-likelihood maximization model	Selected alleles from GISAID genomes	Allele-frequency>=80 % with phylogenetic verification	GitHub or; Via LCS	[34]
Lineagespot	bioconductor	R	VCF	Average allele frequency	Curated mutation profiles	Outbreak.Info or Pangolin	na	[35]
VLQ	na	kallisto, samtools, minimap2, bwa, bbmap Python Packages: pyvcf, pysam	FASTQ	Pseudoalignment to reference genomes (Kallisto)	Selected alleles from GISAID genomes	Reference genomes selected to capture alleles with frequency>=50 %	Via VLQ	[37]
V-pipe	Conda	COJAC, ShoRAH, LolliPop Python Packages: pysam, pandas, numpy, pyyaml, strictyaml, requests, click, poetry-core	FASTQ	Linear model: constrained linear model (Lollipop) and Dirichlet process mixture model (ShoRAH)	Curated mutation profiles	Cov-Spectrum and UKHSA Genomics Public Health analysis variant-definitions	Via COJAC	39–42
Pipes	na	bowtie2, R (ape)	FASTQ	Expectation-maximization algorithm	Selected alleles from GISAID genomes	Calculated phylogenetic internal nodes for each lineage	Via Pipes	[36]