DNAnexus vs Pepkio: Bioinformatics Service Comparison

Quick Comparison Table

What is DNAnexus?

DNAnexus is an enterprise cloud bioinformatics platform hosted on Amazon Web Services (AWS) and Microsoft Azure. It offers security compliance frameworks, including HIPAA, CLIA, CAP, SOC 2, FedRAMP, GDPR, and GxP. The platform serves as the underlying compute infrastructure for large population genomics initiatives, including the UK Biobank Research Analysis Platform (UKB-RAP), as well as clinical diagnostic pipelines and pharmaceutical research.

Users interact with DNAnexus through a web browser interface for pre-built applets or developer interfaces including the dx-toolkit command-line tool, Python/R SDKs, and a REST API. The platform supports workflow languages such as WDL (compiled via dxCompiler), Nextflow (orchestrated for nf-core), and CWL. Supported tools cover bulk RNA-seq, single-cell RNA-seq (Cell Ranger, Seurat, Scanpy), DNA variant calling (GATK, DeepVariant, Sentieon), epigenomics (ChIP-seq, ATAC-seq, bisulfite sequencing), and cohort analytics using Apache Spark, Hail, PLINK, and Regenie.

What does Pepkio offer?

Pepkio provides an outsourced bioinformatics model that handles the entire analytical workflow for research teams. Instead of requiring labs to configure cloud environments, write scripts, or debug pipelines, Pepkio provides direct access to experienced bioinformatics specialists.

Researchers share their raw data files and experimental goals with Pepkio bioinformaticians, who perform quality control, sequence alignment, custom statistical modeling, differential expression, and pathway enrichment. Deliverables include publication-ready vector figures, summary reports, executable script archives (with optional Nextflow or Snakemake workflows and Conda or Docker environments), and written Materials and Methods paragraphs. Pepkio also provides scientist-to-scientist support during peer review to run requested re-analyses and draft responses to reviewer comments.

Pepkio vs DNAnexus: Head-to-Head Comparison

Setup and learning curve

DNAnexus requires technical onboarding, while Pepkio generally does not require local software setup. Running standard GUI applets on DNAnexus can take less than an hour to start, but custom pipelines usually require dx-toolkit (pip install dxpy), API tokens, dxapp.json app specs, Docker containers, or WDL compilation with dxCompiler. For custom runs, teams often spend 1 to 3 days on setup before execution is stable.

With Pepkio, projects usually start with scope discussion and raw data transfer. The lab does not need to manage CLI setup, workflow tooling, or cloud VM configuration.

Analysis depth and customisation

Pepkio provides tailored statistical modeling for complex study designs. DNAnexus provides cloud infrastructure that your team configures and extends. For non-model organisms, non-standard single-cell chemistries, or multi-factor interaction models, DNAnexus often requires custom scripts, Docker builds, or Swiss Army Knife (app-swiss-army-knife) configuration.

With Pepkio, bioinformaticians handle custom reference indexing, de novo genome assembly, nested linear models, and multi-omics integration as part of project delivery.

Time to publishable results

DNAnexus can run core cloud computations quickly. Alignment and variant calling may finish in hours, but getting to publication-ready figures still requires in-house QC review, statistical coding, enrichment analysis, and figure preparation. That downstream work can add substantial time.

Pepkio typically delivers a complete analysis package in 1 to 3 weeks, including figure panels, statistical summaries, and written methods.

Reproducibility and provenance tracking

DNAnexus provides automated platform-level lineage tracking. Every execution receives a unique job ID (job-xxxx) that records exact parameters, software commit hashes, asset IDs, and virtual machine hardware configurations within the project audit log.

Pepkio ensures reproducibility by delivering complete, executable R or Python script archives, raw and processed data tables, and structured Materials and Methods paragraphs detailing exact algorithmic parameters. Workflows in Nextflow or Snakemake, as well as Conda or Docker environments, are available as optional deliverables upon request.

True cost

DNAnexus uses a layered cloud cost model, while Pepkio uses project-based service pricing. DNAnexus costs can include enterprise licensing, storage billed per GB per month, compute runtime, and data egress for downloads. Custom workflows can also require in-house developer hours. Failed jobs from memory sizing errors can still generate charges.

Pepkio charges a single project service fee (exact pricing is not publicly specified), which bundles compute, storage overhead, and service labor.

Troubleshooting and support

DNAnexus offers platform technical support, while Pepkio provides scientist-to-scientist analytical troubleshooting. When jobs fail on DNAnexus, teams usually inspect worker logs, access VMs, or file support tickets. Dependency conflicts and out-of-memory failures still need user-side remediation.

Pepkio handles pipeline maintenance internally. If sample anomalies or execution issues appear, Pepkio bioinformaticians address them as part of the service.

Publication support

Pepkio includes manuscript and peer-review support, while DNAnexus focuses on platform outputs. DNAnexus provides raw outputs (BAM, VCF, count matrices) and HTML QC reports, but figure design and manuscript drafting remain with the research team.

Pepkio delivers custom vector figure panels, drafts Materials and Methods text, and supports peer review through requested re-analyses, threshold adjustments, and response drafting help.

Scaling up

DNAnexus is strong for standardized cohort compute at very large scale. It can run uniform WGS/WES or GWAS pipelines across very large datasets, including UK Biobank-scale cohorts, using parallel nodes and Spark or Hail. Expanding to unfamiliar omics types may require new applets or pipeline work.

Pepkio scales by adding analytical service capacity. Labs can add new analysis types, such as single-cell RNA-seq or epigenomics, without building new internal cloud infrastructure.

Data handling and security

Both options support secure data management under different execution models. DNAnexus provides a certified cloud perimeter compliant with HIPAA, CLIA, CAP, SOC 2, FedRAMP, GDPR, and GxP standards. This makes it suitable for diagnostic laboratories and international consortia that require strict role-based access control within a unified cloud environment.

Pepkio maintains strict data handling protocols throughout the project lifecycle. Raw sequencing data is processed within secure compute environments, and final deliverables are returned directly to the research team.

When to Use Pepkio (Outsource the Analysis)

• Your lab does not have dedicated bioinformaticians and needs managed statistical analysis.
• You have publication deadlines and need figure panels plus drafted Materials and Methods in about 1 to 3 weeks.
• Your study includes non-standard designs, non-model organisms, or custom single-cell chemistries.
• You want direct reviewer-response support for re-analyses and figure updates.
• You prefer fixed service pricing over variable cloud compute, storage, and egress costs.

When to Use DNAnexus (Run It Yourself)

• You are running large-scale biobank or population genetics cohorts, such as UK Biobank-style analyses using Spark, Hail, PLINK, or Regenie.
• You need a certified cloud environment that aligns with HIPAA, CLIA, CAP, SOC 2, FedRAMP, GDPR, or GxP requirements.
• Your team has dedicated bioinformatics developers comfortable with Linux, Python, Docker, and WDL or Nextflow.
• You run multi-center studies that need role-based access and shared standardized workflows.
• You process high volumes of standardized WGS/WES or bulk RNA-seq with established production pipelines.

Trade-Offs at a Glance

Pepkio (Outsourced CRO Model)

• Pros:
  • No local compute setup is typically required.
  • End-to-end delivery includes vector figures and summary reporting in 1 to 3 weeks.
  • Bioinformaticians handle custom contrasts, non-model organisms, and edge cases.
  • Includes support for reviewer-requested re-analyses.
  • Project pricing is generally more predictable than usage-based cloud billing.
• Cons:
  • Up-front per-project costs may be higher than raw cloud compute alone.
  • Pipeline execution is managed by the service team, not internal staff.

DNAnexus (DIY Cloud Platform)

• Pros:
  • Enterprise cloud scalability for very large genomic cohorts (for example, UK Biobank-scale workloads).
  • Compliance frameworks include HIPAA, CLIA, CAP, SOC 2, FedRAMP, GDPR, and GxP.
  • Automated lineage tracking with unique job IDs (job-xxxx) and version hashes.
  • Supports both web GUI use and developer workflows (dx CLI, Python/R SDKs, WDL/Nextflow).
• Cons:
  • Requires in-house bioinformatics and cloud engineering skills.
  • Costs can grow through storage, failed runs, and data egress.
  • Outputs are mainly raw matrices and QC logs, so publication figures and writing remain manual tasks.
  • Custom applet development (dxapp.json) and VM debugging add engineering overhead.

Frequently Asked Questions

Can I still get the underlying code and scripts if I outsource to Pepkio?

Yes. Pepkio provides complete, executable R and Python script archives alongside processed data tables and figure panels upon request. Workflows in Nextflow or Snakemake, as well as Conda or Docker environments, can also be delivered as optional project outputs.

How long does it take to learn DNAnexus for standard RNA-seq analysis?

Using pre-built GUI applets on DNAnexus takes about 15 to 30 minutes to set up a project, upload sample files, and launch a standard pipeline. However, if you need to build custom pipelines, containerize tools in Docker, or write WDL code, expect a learning curve of 1 to 3 days for an experienced bioinformatician.

What happens if a manuscript reviewer asks for a different normalization or statistical threshold?

If you use Pepkio, the service team directly assists with peer-review revisions by re-running statistical models, updating figure panels, and helping draft written responses to reviewers. If you use DNAnexus, your lab must manually adjust parameters, re-run jobs on the cloud platform, and update all downstream figures and text independently.

Does Pepkio require my lab to have cloud computing infrastructure?

No. Pepkio operates as a full-service provider, handling all computational processing on its own infrastructure. Your lab does not need local HPC servers, cloud accounts, or specialized software installations.

Can DNAnexus run custom Nextflow or WDL pipelines?

Yes. DNAnexus supports WDL compiled via dxCompiler and Nextflow workflows running on head-node workers (including standard nf-core pipelines). Developers must configure app specifications (dxapp.json) and allocate appropriate instance resources.

How are cloud storage and compute costs handled on DNAnexus?

DNAnexus bills compute based on cloud instance runtimes (AWS EC2 or Azure VMs plus platform markup) and charges monthly storage fees per gigabyte for active and archival data. Additional charges apply when downloading files off the platform via data egress fees.

What bioinformatics skills are needed to use DNAnexus effectively?

Running basic pre-built apps via the DNAnexus web GUI requires basic domain knowledge. However, developing custom applets, troubleshooting failed runs, or building cohort analytics pipelines requires proficiency in the Linux command line, Python or Bash scripting, Docker containerization, and WDL or Nextflow.

Is Pepkio suitable for non-model organism genomics?

Yes. Pepkio's bioinformaticians regularly handle non-model organisms, including de novo genome assembly, custom annotation, and non-standard reference indexing. On DIY platforms like DNAnexus, non-model organism analyses require manual configuration by the user.

How does reproducibility compare between Pepkio and DNAnexus?

DNAnexus provides automated platform-level lineage, logging exact job execution IDs (job-xxxx), software hashes, and virtual machine hardware specs for every run. Pepkio delivers reproducibility through documented Materials & Methods text, parameter records, data matrices, and executable script archives (with optional container environments).

Can DNAnexus handle single-cell RNA-seq datasets?

Yes. DNAnexus supports single-cell RNA-seq processing through integrated tools like 10x Genomics Cell Ranger, Seurat, and Scanpy, as well as interactive JupyterLab and RStudio notebooks. Users are responsible for parameter tuning, cell filtering, and cluster annotation.

What is the main cause of unexpected costs on DNAnexus?

Unexpected costs on DNAnexus typically stem from accumulating uncompressed intermediate BAM or FASTQ files in active cloud storage, downloading large datasets (data egress fees), or running long compute jobs that fail due to virtual machine memory sizing errors.

Who writes the Materials and Methods section for publication?

When working with Pepkio, the service team drafts complete, publication-ready Materials and Methods paragraphs detailing the exact algorithms, versions, and statistical tests used. When using DNAnexus, the researcher must write the Methods section based on job execution logs.

Bottom Line

Choosing between Pepkio and DNAnexus depends on whether your team needs cloud infrastructure for internal pipeline engineering or managed scientific deliverables for a project. DNAnexus is often the better fit for biobanks, enterprise pharma teams, and consortia with dedicated bioinformatics developers and strict cloud governance needs. Pepkio is often the better fit for labs that want outsourced analysis, publication-ready figures, statistical support, and reviewer-response assistance without building internal pipeline operations.