From Mexico’s First Sequencer to a Global Roadmap: Making Transcriptomics FAIR and Accessible

Seventeen years ago, the arrival of Mexico’s first Illumina sequencing system at the National Autonomous University of Mexico (UNAM) marked the beginning of a quiet revolution in the country’s life sciences. Now, in a comprehensive review published in Frontiers in Bioinformatics, Verónica Jiménez Jacinto of the UNAM Institute of Biotechnology and Leticia Vega-Alvarado of the UNAM Institute of Applied Sciences and Technology have distilled more than 17 years of transcriptomic experience into a strategic roadmap for RNA sequencing analysis. Their article, which traces the entire journey from RNA extraction to biological interpretation, arrives at a moment when the field faces a dual challenge: ensuring that analyses are technically sound and reproducible, and breaking down the barriers that keep powerful bioinformatics tools out of reach for many researchers.

Whole-transcriptome analysis, which examines the total RNA within a biological sample, has become a cornerstone of contemporary molecular biology because it allows the expression of both coding and non-coding genes to be characterized in a single experiment. The applications span remarkable ground. In agricultural and animal sciences, transcriptomics has advanced veterinary parasitology, teleost physiology, and plant-pathogen interactions. In genomic medicine, it illuminates pathological mechanisms; in pharmacology, it supports molecular-level studies across organisms; and in structural biology, it reveals the functions of small RNAs during transcription. Yet the authors caution that every step of the workflow, from initial data processing to final interpretation, can introduce biases that substantially affect biological conclusions.

The review organizes the field around three major scenarios that determine which analytical strategies are appropriate. When a well-annotated reference genome exists, researchers can align short sequencing reads to the genome and quantify gene abundance with relative ease. When only a closely related annotated organism is available, transcriptome assembly becomes more involved. And for non-model species lacking any adequate reference, de novo assembly is required—a far more demanding task. The choice of workflow also depends on the class of RNA under study, whether messenger RNA, small regulatory RNAs such as microRNAs, or long non-coding RNAs, each of which demands distinct library preparation and computational treatment.

Experimental design receives particular scrutiny in the review, and the authors’ practical recommendations challenge some long-standing conventions. While three biological replicates have traditionally served as an accepted minimum for differential expression analysis, the authors note that this number may be insufficient to capture biological variability or detect small-magnitude expression changes. Citing published work showing that increasing replicates to between six and twelve can substantially improve detection of differentially expressed genes, they recommend that six or more replicates may provide a useful balance between experimental cost and statistical power, particularly for studies involving patient samples or outbred animal models with substantial inter-individual variability. Homogeneous materials such as cell lines, by contrast, can yield satisfactory results with at least three replicates when expression changes exceed twofold.

Batch effects emerge as one of the most insidious threats to transcriptomic validity. This form of technical variation arises when samples are processed in different batches, on different days, in different laboratories, or by different technicians, and it can generate false positives—genes flagged as differentially expressed when the difference actually reflects the day they were sequenced. Drawing on their experience at the University Unit for High-Throughput Sequencing and Bioinformatics, the authors identify RNA extraction as a particularly sensitive stage, where small differences in processing time, temperature, and reagent lots can introduce systematic variation. Exploratory analyses such as principal component analysis and multidimensional scaling, performed on normalized expression data, can reveal clustering patterns associated with processing batches rather than biological conditions, and batch can then be included as a covariate in statistical frameworks such as DESeq2 or edgeR.

On the sequencing side, the review weighs the trade-offs between established short-read platforms and emerging long-read technologies. Illumina systems remain widely used because of their low error rates, high throughput, and relatively low cost per base, making them cost-effective for studies involving large numbers of samples. For de novo transcriptome assembly, however, long-read sequencing from Oxford Nanopore Technologies and Pacific Biosciences offers important advantages: reads that span much larger portions of, or in some cases entire, transcripts simplify reconstruction of isoforms and reduce uncertainty about exon connectivity. PacBio HiFi sequencing achieves very high single-read accuracy, while ONT supports direct RNA sequencing without cDNA conversion, reducing biases from reverse transcription and PCR amplification and enabling detection of RNA base modifications. The higher cost of long-read approaches remains a limiting factor.

The computational pipeline itself is mapped in careful technical detail. Quality control begins with tools such as FastQC and MultiQC to assess raw sequencing quality across samples, followed by preprocessing with Cutadapt, fastp, or Trimmomatic to remove adapters and low-quality bases. For de novo assembly from short reads, de Bruijn graph-based assemblers Trinity and rnaSPAdes reconstruct transcripts through multistage processes, while RNA-Bloom2, isONform, and RATTLE serve long-read data. Assembly quality can then be evaluated with reference-free tools such as TransRate or reference-based suites including rnaQUAST and DETONATE. When a reference genome is available, splice-aware aligners HISAT2 and STAR map reads while identifying splice junctions, and quantification tools ranging from Salmon and Kallisto to RSEM, StringTie, HTSeq, and featureCounts convert alignments into abundance matrices. For long-read data, Bambu, IsoQuant, and Oarfish provide transcript and isoform quantification.

Differential expression analysis, the statistical heart of most transcriptomic studies, is examined through the leading frameworks. DESeq2 and edgeR both use negative binomial models to account for biological variability and overdispersion in RNA-seq count data, with edgeR employing an empirical Bayes approach to dispersion estimation that shares information across genes. The limma framework, adapted for RNA-seq through the voom transformation, uses linear models on log-transformed expression values with precision weights, offering computational efficiency for complex designs. The non-parametric NOISeq estimates noise distributions empirically and may be useful when distributional assumptions are difficult to satisfy, though the authors stress that the absence of biological replicates substantially limits statistical inference regardless of method. The resulting gene lists then feed into functional enrichment analysis against databases such as Gene Ontology, KEGG, and COG, using approaches from over-representation analysis to gene set enrichment analysis.

Woven through the technical guidance is a sustained argument for FAIR compliance—the principles that data should be Findable, Accessible, Interoperable, and Reusable. The authors detail what this means in practice: persistent identifiers such as GEO or BioProject accession numbers and stable gene identifiers rather than gene names alone; deposition of raw FASTQ files in public repositories such as the NCBI Sequence Read Archive, the European Nucleotide Archive, or the DNA Data Bank of Japan; standardized formats such as BAM, GTF/GFF3, and CSV; and comprehensive documentation of software versions, parameter settings, and reference genome assemblies. Containerization tools such as Docker and Conda can preserve the computational environment, strengthening reproducibility of quantitative results.

The review culminates in a call for bioinformatics democratization, embodied in IDEAMEX, a web-based platform the authors developed to facilitate transcriptomic analysis from abundance matrices. The platform supports exploratory data analysis and differential expression analysis using four methods—DESeq2, edgeR, limma, and NOISeq—whose integration provides complementary evidence for identifying robust expression patterns, all without requiring advanced programming skills. The authors frame this not merely as a technical convenience but as a response to systemic inequalities: women remain underrepresented in bioinformatics in Mexico as in other STEM disciplines, and researchers in resource-limited settings often lack access to essential computing infrastructure, mentoring, and training. Combined with FAIR-compliant workflows, they argue, such resources can contribute to more open, reproducible, and inclusive access to genomic knowledge, regardless of available computational infrastructure or financial resources.

Subject of Research: FAIR-compliant transcriptomic workflows and democratization of RNA-seq bioinformatics

Article Title: Strategic roadmap for FAIR-compliant transcriptomic workflows: from technical challenges to bioinformatics democratization

Article References: Jiménez Jacinto, V., & Vega-Alvarado, L. (2026). Strategic roadmap for FAIR-compliant transcriptomic workflows: from technical challenges to bioinformatics democratization. Frontiers in Bioinformatics, 6, Article 1882808. https://doi.org/10.3389/fbinf.2026.1882808

Image Credits: AI Generated

DOI: 10.3389/fbinf.2026.1882808

Keywords: transcriptomics, RNA-seq, FAIR principles, bioinformatics, differential expression, de novo assembly, IDEAMEX, UNAM, reproducibility, democratization, Strategic, roadmap