Figure 1: Decision landscape for agricultural transcriptome projects — building in-house capacity versus partnering with an external sequencing and analysis provider.

Running a transcriptome project for crops or livestock involves more than extracting RNA and sending it to a sequencer. Researchers managing multi-sample studies across species, tissues, and treatments face a recurring tension: invest in in-house infrastructure or outsource to a specialized provider. The right answer depends on sample logistics, bioinformatics capacity, and whether the data need to hold up under peer review. This guide walks through the practical tradeoffs — from choosing between mRNA-seq and total RNA-seq to reading a QC report without getting lost in the metrics — so that research teams can decide what to keep in-house and what to hand off.

Key takeaways

Outsourcing bulk RNA-seq makes the most economic and operational sense when a project demands consistent library preparation, sequencing depth, and bioinformatics across 12 or more samples from species without mature in-house pipelines.

The largest cost driver is not the sequencing itself but the bioinformatics — alignment, quantification, differential expression, and functional annotation — especially for non-model crops and livestock species with fragmented reference genomes.

Provider selection hinges on three concrete criteria: demonstrated experience with your species or tissue type, transparent QC reporting with actionable pass/fail thresholds, and deliverable formats that integrate with your downstream analysis tools.

A well-structured data package should include raw FASTQ files, aligned BAM files, expression count matrices, and a QC summary that maps each metric to a documented threshold, not just a "passed" checkbox.

Building It In-House vs. Sending It Out

Setting up an in-house RNA-seq pipeline for agricultural species looks attractive on a grant budget spreadsheet. The sequencer amortizes over hundreds of runs, students and postdocs provide labor, and the PI retains full control over every step. In practice, the hidden costs pile up quickly.

For projects that cross species boundaries — say, drought-stress transcriptomics in sorghum alongside rumen epithelium profiling in cattle — the in-house team must maintain two separate reference genomes, annotation sets, and quantification parameters. An outsourced provider that routinely handles both plant and livestock transcriptomes can absorb that complexity without the PI assembling a cross-disciplinary bioinformatics team. For studies where transcriptome data need to be interpreted alongside epigenomic information — such as stress-responsive gene regulation — the guide on epigenomic sequencing for agricultural research covers the complementary methods that explain expression patterns at the chromatin level.

Research groups considering in-house sequencing for agricultural applications may find useful context in the overview of next-generation sequencing platforms for plant and animal breeding: NGS technology in animal and plant breeding. For groups already running genomic studies, the resource on research development in plant genome sequencing provides additional background on how sequencing strategies have evolved for crop species.

Picking the Right RNA-Seq Strategy

Before approaching any provider, the research question should dictate the RNA-seq method — not the other way around. Agricultural transcriptome projects tend to cluster into three strategy buckets, and picking the wrong one guarantees wasted reads.

mRNA-seq for gene expression

Poly-A enrichment captures mature mRNA and is the default choice for differential expression analysis in most crop and livestock tissues. It works well with high-quality RNA (RIN ≥ 7) and produces cleaner data with less ribosomal contamination. For standard comparisons — treated vs. control in leaf, root, or muscle tissue — a transcriptome sequencing service running mRNA-seq at 20–30 million paired-end reads per sample provides sufficient power for detecting fold changes of 1.5× or greater.

Total RNA-seq when mRNA falls short

Some agricultural samples resist poly-A selection. Degraded RNA from field-collected tissues, formalin-fixed samples, or stress-treated plants with active RNases can lose mRNA during enrichment. Total RNA-seq with ribosomal depletion captures mRNA, long non-coding RNA, and nascent transcripts even when RNA integrity is compromised. The tradeoff: 30–50% more reads are needed because rRNA depletion efficiency varies, and the resulting libraries contain a broader transcript population.

Beyond bulk: when single-cell or spatial methods apply

Single-nucleus RNA-seq is gaining traction in plant research for resolving cell-type-specific responses in complex tissues like root tips, developing embryos, and meristems. Direct RNA sequencing using nanopore platforms eliminates reverse transcription bias and captures RNA modifications — an emerging strategy for plant epitranscriptomics covered in the overview of direct RNA sequencing applications in plant transcriptomics. For projects requiring full-length isoform characterization, long-read sequencing platforms discussed in the de novo genome assembly guide also support transcript-level analysis that complements short-read RNA-seq data.

Strategic checklist

Sample quality drives method choice. RIN ≥ 7 and 500 ng total RNA per sample: mRNA-seq works. RIN < 6 or limited material: total RNA-seq with ribosomal depletion, or request a provider QC consultation before committing.

Species reference quality matters. Well-annotated genomes (rice, maize, soybean, cattle, pig, chicken) support standard STAR → featureCounts → DESeq2 pipelines with minimal customization. Draft or fragmented assemblies for non-model crops and minor livestock breeds add bioinformatics complexity that should be reflected in the provider's pricing and timeline. A bioinformatics analysis service with agricultural genomics experience can handle custom reference preparation and species-specific pipeline optimization that generalist providers typically charge as add-on line items.

Replicate count affects power more than read depth. For agricultural field experiments with high biological variability — common in drought, heat, and pathogen-challenge studies — five or six biological replicates per condition deliver more statistical power than doubling read depth on three replicates.

Questions Every Provider Should Answer

Before sending samples, a brief conversation with the provider's technical team can surface mismatches that derail projects later. The following questions cut through marketing language and get to operational reality.

For researchers designing abiotic stress experiments in crops, the methodology choices covered in the guide on designing RNA-seq experiments for plant abiotic stress can help frame specific technical questions for providers about stress-responsive transcript detection, time-course replicates, and normalization strategies.

What Your Data Package Must Include

A complete outsourced transcriptome project should deliver more than raw sequence files. Researchers who plan to publish need data packages that survive reviewer scrutiny — and a provider with dedicated bioinformatics analysis

experience in agricultural species should structure deliverables accordingly. The following table maps each deliverable to its purpose and the minimum format requirements.

A frequently overlooked item: the count matrix should contain raw integer counts, not normalized values (FPKM, RPKM, TPM). Normalized matrices hide library-size differences that differential expression tools rely on for accurate dispersion estimation. If the provider supplies only normalized counts, ask for the raw matrix — most will provide it on request.

QC Metrics Worth Watching

RNA-seq QC reports can run to dozens of metrics. The five that matter most for agricultural transcriptome projects cover data integrity from library preparation through alignment. A provider's agricultural transcriptomic data analysis team should flag any metric that falls outside documented thresholds before delivering the final data package.

Figure 2: Five core QC metrics for outsourced RNA-seq projects, with typical agricultural-species reference ranges.

Metric 1: Total reads per sample

For standard differential expression in well-annotated crop and livestock genomes, 20–30 million paired-end reads per sample typically provides adequate gene detection sensitivity. Lower counts — below 15 million — risk missing low-abundance transcripts, including transcription factors and stress-responsive genes that are often the targets of agricultural research.

Metric 2: Mapping rate

Expect 80–95% of reads to map uniquely to the reference genome for well-annotated species. Mapping rates below 70% in rice, maize, cattle, or pig suggest issues with sample contamination, poor reference choice, or excessive rRNA carryover. For non-model crops with draft genomes, 60–75% may be acceptable but should be discussed with the provider upfront.

Metric 3: rRNA contamination

Ribosomal RNA should account for less than 5% of reads in poly-A-selected mRNA-seq libraries. Values above 10% indicate inefficient mRNA enrichment, which reduces effective sequencing depth for mRNA targets. For total RNA-seq with ribosomal depletion, residual rRNA below 15% is standard.

Metric 4: Gene assignment rate

The fraction of aligned reads that overlap annotated gene features should fall between 50–70% depending on annotation quality. Low assignment rates with high mapping rates often point to an annotation file that is incomplete or mismatched to the genome assembly — a common issue in agricultural species where annotation updates lag behind assembly improvements.

Metric 5: 3' bias and coverage uniformity

Transcript coverage should be roughly uniform along the gene body. Strong 3' bias (coverage skewed toward the 3' end) indicates RNA degradation during library preparation. This is common in field-collected samples exposed to heat or delayed processing; it does not necessarily invalidate the data but should be documented and, where possible, accounted for in the analysis.

Turnaround Realities and Cost Drivers

Agricultural transcriptome projects rarely follow the tidy timelines advertised on service websites. Seasonality, sample logistics, and reference genome quality all influence how long a project takes and what it costs.

A project with 24 maize leaf samples for mRNA-seq at 25M reads, with a well-annotated B73 reference genome, can reasonably complete in 4–6 weeks from sample receipt to data delivery. The same project for 24 cassava root samples with a fragmented reference adds 2–3 weeks for genome-guided transcriptome assembly and custom annotation.

Please note that the pricing examples and discount thresholds mentioned above are for reference only. CD Genomics evaluates each project individually based on species, sample conditions, experimental design, and analytical depth. For a project-specific quotation, please contact us directly.

Planning Your First Outsourced Project

Figure 3: A five-step project planning checklist for first-time RNA-seq outsourcing, from pre-submission quality control through post-delivery re-analysis negotiation.

First-time outsourcing typically introduces friction that experienced labs have already solved. Anticipating these friction points keeps projects on schedule.

Pre-submission QC saves weeks. Run RNA integrity and concentration on every sample before shipping. A Nanodrop alone is not enough — a Bioanalyzer or TapeStation trace confirms whether the RNA is intact or degraded. Catching one degraded sample before it enters the provider's pipeline avoids a downstream re-extraction and re-submission cycle that can add 3–4 weeks.

Freeze an annotation version before analysis starts. Providers default to the most recent genome assembly and annotation available for your species. If your lab has been working with a specific version for previous studies, specify it explicitly. Switching annotation versions mid-project creates irreproducible gene counts.

Plan for the metadata sheet. Every provider needs a sample metadata table with sample ID, tissue, treatment group, replicate number, and species. Formatting inconsistencies — spaces vs. underscores in group names, missing replicate labels — cause delays that are entirely avoidable. Send the metadata sheet alongside the samples, not two weeks later.

Reserve time for data QC on your end. A provider's QC report confirms the data are technically sound. Whether the data answer your biological question is a separate check. Run a quick PCA or sample-to-sample distance heatmap on the count matrix within 48 hours of receiving it. If samples cluster by batch or technical factors rather than biological groups, flag it immediately.

Negotiate re-analysis windows upfront. Most providers include one round of re-analysis within a defined window (typically 30–60 days post-delivery). Confirm whether this covers parameter changes, alternative reference genomes, or additional normalization methods — and whether it resets if you discover an issue late.

FAQ

Q1: When does outsourcing RNA-seq cost less than running it in-house?

A: The crossover point depends less on sequencer amortization schedules than on bioinformatics labor. When a project involves non-model species requiring custom reference preparation, or when the lab lacks a dedicated bioinformatician who can maintain species-specific pipelines across multiple projects, outsourcing typically becomes cheaper once annual sample volume falls below 200–300 samples. Above that threshold, in-house sequencing may cost less per sample, but the bioinformatics personnel cost remains independent of sample count and should be factored separately.

Q2: How many biological replicates do I need for an agricultural field experiment?

A: For field-based stress experiments where environmental variability is high — drought, heat, pathogen infection in open-field conditions — five to six biological replicates per treatment group provide more reliable differential expression calls than three replicates with deeper sequencing. Three replicates may be adequate for controlled-environment growth chamber studies, but the added variability of field conditions, soil heterogeneity, and microclimate effects nearly always justifies the extra replicates. Power analysis tools such as PROPER or RNASeqPower can estimate required sample sizes using pilot data or published coefficient-of-variation estimates for your species and tissue.

Q3: What reference genome should my provider use for alignment?

A: The provider should use a published, versioned reference assembly — not an internally modified draft — and must state the version explicitly in the methods and QC report. For major crops, current standards include IRGSP-1.0 for rice, B73 RefGen_v4 or v5 for maize, and Glycine_max_v2.1 for soybean. For livestock, ARS-UCD1.2 for cattle, Sscrofa11.1 for pig, and GRCg6a for chicken are widely accepted. If your species has multiple published assemblies, discuss with the provider which one aligns with your publication venue's expectations before analysis begins.

Q4: Can I submit samples from different species or tissue types in the same project batch?

A: Yes, and most agricultural transcriptome providers routinely handle mixed-species projects. The key requirement is that the provider maps each sample to its correct reference genome and annotation version, reports QC metrics per species rather than pooled across species, and applies appropriate batch-correction steps if samples from different species are sequenced on the same flow cell. Confirm these details during project design rather than assuming they are handled automatically.

Q5: What should I check when I first receive my data package?

A: Within 48 hours of receiving the data, run a principal component analysis on the raw count matrix and check whether samples cluster by the biological groups you expect. If samples cluster by sequencing batch, library preparation date, or an unlabeled technical factor instead of by treatment group, contact the provider immediately. Also verify that the number of FASTQ files matches your sample count, that all BAM files are indexed and readable, and that every sample in your metadata sheet appears in the count matrix with the correct identifiers.

How CD Genomics Can Help

As an agricultural genomics provider, CD Genomics supports transcriptome projects across crop and livestock species, from study design through sequencing and bioinformatics delivery. The transcriptome RNA-seq services cover bulk mRNA-seq, total RNA-seq with ribosomal depletion, small RNA-seq, and direct RNA sequencing on both Illumina and nanopore platforms. For projects requiring deeper interpretation, the agricultural transcriptomic data analysis team provides differential expression, pathway enrichment, co-expression network analysis, and custom visualization. A broader overview of available platforms is available on the agricultural genomics services overview. All laboratory services described here are intended for Research Use Only.

Research Use Only Statement

The information provided in this article is for research use only and is not intended for use in diagnostic or therapeutic procedures. CD Genomics provides sequencing and bioinformatics services for research purposes. Researchers should consult the appropriate regulatory guidelines for their specific applications.

References

1. Conesa A, Madrigal P, Tarazona S, et al.. "A survey of best practices for RNA-seq data analysis." Genome Biology, 2016;17:13.. doi:10.1186/s13059-016-0881-8
2. Stark R, Grzelak M, Hadfield J.. "RNA sequencing: the teenage years." Nature Reviews Genetics, 2019;20(11):631-656.. doi:10.1038/s41576-019-0150-2
3. Van den Berge K, Hembach KM, Soneson C, et al.. "RNA sequencing data: hitchhiker's guide to expression analysis." Annual Review of Biomedical Data Science, 2019;2:139-173.. doi:10.1146/annurev-biodatasci-072018-021255
4. Love MI, Huber W, Anders S.. "Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2." Genome Biology, 2014;15(12):550.. doi:10.1186/s13059-014-0550-8