Single-cell RNA-seq analysis

We routinely analyse data from 10x Genomics Chromium (3' and 5') and Parse Biosciences Evercode, as well as BD Rhapsody, Smart-seq2/3 and other droplet or combinatorial-barcoding chemistries. We can start from the manufacturer's raw output (Cell Ranger, split-pipe) or from a count matrix you have already generated.

Either. We can process raw FASTQ or BCL files through quantification (Cell Ranger, STARsolo, salmon alevin-fry or kallisto bustools, and split-pipe for Parse data), or start from an existing count matrix or object (10x .h5, .h5ad / AnnData, Seurat .rds) if quantification has already been done.

Batch effects are addressed first at the design level, then computationally with methods such as Harmony, scVI/scANVI or Seurat RPCA, chosen to fit your experimental structure. We always check that integration preserves genuine biological signal rather than erasing it, because some batch confounding cannot be corrected after the fact. We cover this in detail in our article on batch effects in scRNA-seq.

We combine three approaches: manual annotation from canonical marker genes, reference-based annotation against curated atlases (SingleR, Azimuth, CellTypist), and AI-assisted annotation to accelerate and cross-check labels. Final labels are reviewed by an expert and justified by marker evidence, not accepted blindly from an automated call.

Yes. For comparisons between conditions or groups we use a pseudo-bulk approach (aggregating counts per sample and cell type, then modelling with DESeq2 or edgeR), which controls false positives far better than naive per-cell tests. We also provide per-cluster marker genes and, where relevant, compositional analysis of cell type proportions.

Yes. We infer differentiation trajectories and pseudotime with tools such as Slingshot, Monocle3 and PAGA, and can add RNA velocity (scVelo) to indicate directionality. These analyses are applied when they answer a biological question, not as decoration.

Yes. We process paired single-cell gene expression and VDJ data (10x 5' immune profiling and equivalents), characterise TCR/BCR clonotypes, clonal expansion and diversity, and link repertoire features back to transcriptional cell states. This is a particular area of expertise.

Yes. We integrate scRNA-seq with CITE-seq surface proteins, single-cell ATAC (multiome), spatial transcriptomics and matched bulk RNA-seq, using multimodal methods such as totalVI, MultiVI and WNN to build a coherent picture across layers.

We sign a non-disclosure agreement before any data transfer, exchange data through secure channels and never share or reuse your data. Where your institution has specific security requirements, we comply with them fully.

The cost depends on the platform, the number of samples and cells, and the depth of downstream analysis requested (annotation only, or with differential expression, trajectories, repertoire and multi-omics integration). Every project is scoped and quoted transparently before any work begins. Send us your design and data type and we will reply with a clear scope, timeline and flat-rate quote within 48 hours.