CD Genomics

09/02/2026

🧬 The HLA region is the most polymorphic part of the human genome — and your typing strategy decides what your data is worth.

Ambiguous calls, phase unknowns, and gene conversions between neighboring loci are everyday problems in HLA research. Choosing the right platform matters:

✔ Sanger sequencing — reliable resolution for well-characterized loci ✔ NGS — scalable multi-locus typing for cohort studies ✔ PacBio and Nanopore long reads — full-gene haplotypes where short reads leave phase ambiguous

Our HLA typing service covers class I and II genes (HLA-A, -B, -C, -DR, -DP, -DQ, plus HLA-E and -G) across all of these platforms — so the method fits your study design, not the reverse.

We also offer KIR typing: 14 KIR genes plus 2 pseudogenes, allele-level resolution, and copy number variation analysis — a useful addition for NK cell and innate immunity studies.

Typing panels can be scoped to 6, 11, or 15 loci, with clear sample requirements per input type.

📄 Explore the service: https://www.cd-genomics.com/hla-typing.html

09/02/2026

🧬 From patient neoantigen to a candidate T cell receptor — how do you bridge that gap?

Hepatocellular carcinoma has a suppressive immune microenvironment and limited T cell infiltration, which has held back T cell-based immunotherapy development.

A study in Gut (Maravelia et al., 2025) took a neoantigen-driven approach:

→ Screened patient-specific neoantigens through co-culture → Isolated neoantigen-reactive T cells expressing 4-1BB → Characterized these cells through T cell receptor sequencing

For this study, the authors used bulk TCR sequencing provided by CD Genomics to resolve the receptor repertoires of the isolated tumor-reactive T cell populations.

The team then reconstructed the recovered TCRs and functionally evaluated them — building a complete pipeline from patient neoantigen identification to candidate receptors for T cell-based immunotherapy.

It's a good example of repertoire sequencing doing real work in translational immunology: not just describing the immune response, but enabling the next step.

📄 Read the full paper:
https://doi.org/10.1136/gutjnl-2024-334148

09/02/2026

🤖 Most single-cell foundation models only "see" the 1,500–2,000 most highly expressed genes. A new model published in Nature Communications sees all ~28,000.

scLong (Bai et al., 2026) is a billion-parameter foundation model for single-cell transcriptomics built around two ideas:

🧬 Genome-wide self-attention — the model attends across all 27,874 human genes per cell, including the low-expression and zero-expression genes that other models drop (and where many transcription factors and receptors live)

📊 Gene ontology integration — a graph convolutional network folds GO annotations into the gene embeddings, adding biological context

The model was pretrained on 48 million cells from 1,618 scRNA-seq datasets across 50+ tissues.

In benchmarks, it outperformed existing foundation models at:

→ Predicting transcriptional responses to genetic and chemical perturbations → Forecasting cancer drug responses, including combinations → Inferring gene regulatory networks → Zero-shot batch integration

The takeaway for single-cell researchers: the excluded low-expression genes may hold the regulatory information your current models are missing.

📄 Read the paper: https://doi.org/10.1038/s41467-026-69102-y

Here the authors present scLong, a billion parameter foundation model trained on 48M single cells, using self attention across all 28k human genes and Gene Ontology knowledge. It captures long range gene context in single Cell transcriptomics.

09/02/2026

🧬 Flow cytometry tells you T cells expanded. Sequencing tells you which receptors did it.

Immune repertoire sequencing reveals the specific TCR and BCR clonotypes that expand, contract, and persist in response to tumors, pathogens, vaccines, or autoimmune triggers — information phenotype-based methods simply can't provide.

What our TCR & BCR sequencing service offers:

✔ 5'RACE-based library prep — captures the complete V(D)J region without the primer bias of multiplex PCR ✔ All TCR chains (TRA, TRB, TRG, TRD) and BCR chains (IGH, IGK, IGL) ✔ Input from RNA, PBMC, whole blood, tissue, gDNA, and FFPE samples ✔ IMGT-aligned gene usage, clonotype annotation, and diversity metrics (Shannon, Simpson, clonality) ✔ V-J pairing and CDR3 length distribution for clonal tracking

From immunotherapy monitoring and vaccine studies to autoimmunity and transplantation research — repertoire data adds a dimension your phenotype panels can't reach.

📄 Explore the service: https://www.cd-genomics.com/tcr-seq.html

08/25/2026

DNA methylation is programmed — and the genetic sequence itself determines where and when. Rosenski, Sabag et al. (April 2026, Nature Communications) used WGBS across 39 human cell types and in**ed mouse strains to map how genetic variation shapes methylation.

Key findings: thousands of meQTLs (methylation QTLs), two programming windows (implantation and organogenesis), 33,574 allele-specific methylation regions controlled by common SNPs, overlapping with eQTLs, enhancers, and disease variants.

Why it matters: interpreting DNA methylation changes without knowing the underlying genotype can misattribute cause and effect. A methylation difference might be genetic, not environmental.

Cross-species resource for interpreting WGBS data and designing epigenome-wide association studies.

Open access: https://doi.org/10.1038/s41467-026-71693-5

Genetic influences on DNA methylation across development are poorly understood. Here, the authors map sequence‑dependent methylation in mouse and human, identifying developmental windows and 33,574 human regions where SNPs control allele‑specific methylation linked to regulatory and disease loci...

08/25/2026

Predicting 3D genome organization from ATAC-seq alone — no Hi-C library needed. Hi-Compass (Nature Communications, April 2026) is a depth-aware deep learning framework that predicts cell-type-specific chromatin contacts from chromatin accessibility data.

Key innovation: a depth-aware module that adapts to sequencing depth from sparse single-cell to high-coverage bulk ATAC-seq. Outperformed Akita, C.Origami, Epiphany, and ChromaFold. Zero-shot prediction on unseen cell types. Cross-species transfer to mouse.

Practical impact: existing scATAC-seq atlases can now be mined for 3D genome structure without additional Hi-C library prep. GWAS noncoding variants connected to their 3D regulatory context.

Open access: https://doi.org/10.1038/s41467-026-71877-z

08/25/2026

Invasive species management: did the population arrive once or multiple times? A 2025 Molecular Ecology study used 2b-RAD sequencing services provided by CD Genomics to trace the invasion history of the German wasp (Vespula germanica).

2b-RAD genotyping generated thousands of genome-wide SNPs from specimens across introduced and native ranges. Results: genomic evidence for dual introduction events, reduced diversity in introduced populations (founder events), and genetic admixture between the two introduced lineages.

2b-RAD delivers 10,000+ SNPs across hundreds of individuals at a fraction of whole-genome sequencing cost. Ideal for population structure, migration patterns, bottleneck detection, and hybrid zone mapping.

Our 2b-RAD services support population genetics across all taxa — insects, plants, livestock, fish, microbes.

Open access: https://doi.org/10.1111/mec.70217

Biological invasions are major drivers of recent biodiversity changes, yet the genetic structure and ecological mechanisms underlying invasion dynamics remain poorly resolved in invasive social insec...

08/25/2026

ATAC-seq reveals which genomic regions are open for transcription factor binding — and which are packed away. Our ATAC-Seq Service profiles open chromatin genome-wide from as few as 500 cells.

Tn5 transposase simultaneously fragments DNA and tags adapters. No sonication. No ligation bias. Minimal input.

Deliverables: active regulatory regions (promoters, enhancers), TF footprinting at base-pair resolution, cell-type-specific accessibility changes, nucleosome positioning. Integration with RNA-seq and ChIP-seq.

500-50,000 cells for scATAC-seq, 50,000+ for bulk. Human, mouse, plant, non-model species. 10x and BD Rhapsody scATAC-seq available.

Profile your regulatory landscape: https://www.cd-genomics.com/epigenetics/atac-seq.html

08/25/2026

Genome sequence tells you what genes are there. Hi-C tells you how they are organized in 3D space — and that organization controls which genes are turned on.

Our Hi-C Service maps chromatin interactions genome-wide: TADs, compartment domains, enhancer-promoter contacts. Restriction enzyme digestion + proximity ligation + Illumina NovaSeq sequencing. Bioinformatics deliverables include contact matrices, TAD calling, differential loop analysis, and A/B compartment scoring.

100,000-1,000,000 cells. Fresh or flash-frozen tissue (min 20 mg). Human, mouse, plant, non-model species. .hic and cooler output formats.

Chromosome architecture, structural variant breakpoints, cell-type-specific reorganization — all from one Hi-C library.

Map your genome's 3D architecture: https://www.cd-genomics.com/epigenetics/hi-c-service.html

08/19/2026

Single-cell RNA-seq tells you which cells are in a tumor. Long-read single-cell RNA-seq reveals which isoforms those cells use — and for glioblastoma, that difference matters.

A new Nature Communications study analyzed 182,222 cells from 27 GBM patients using single-cell long-read RNA sequencing. Key findings: → 6,524 unannotated isoforms discovered, 179 tumor-specific → 40 "surface-intracellular" target pairs identified for bispecific therapy → Novel tumor neoantigen candidates for personalized vaccines → Hundreds of isoforms differentially used across tumor subpopulations

This is a new dimension of tumor heterogeneity that short-read sequencing cannot detect.

Read the full paper: https://doi.org/10.1038/s41467-026-72258-2

Short-read sequencing-based approaches are limited to gene-level expression and cannot identify full-length transcript isoforms. Here, the authors apply single-cell long-read RNA sequencing to glioblastoma and identify tumor-specific isoforms and peptides deriving from them as neoantigens.