SeekGene Platform Reference Genome Guide
To ensure high-quality and reliable analysis of single-cell and multi-omics data, the SeekGene cloud platform not only integrates standardized reference genomes released by internationally recognized databases, but also performs in-depth customization and optimized filtering of the underlying annotation data based on the characteristics of single-cell and spatial sequencing technologies. This helps minimize background noise and improve the capture of biologically meaningful signals.
This document provides an overview of the reference genomes available on the platform for human, mouse, and rat, including assembly versions, chromosome composition, and gene type statistics. It also summarizes common download and naming rules for reference genomes, making it easier for users to verify files or filter relevant genes by prefix in downstream custom analyses.
1. Built-in Reference Genomes on the Platform
1.1 Human
- Genome assembly: GRCh38
- Annotation version: GENCODE v44
- Original download link: GENCODE v44 GTF
Chromosome and gene statistics
- Chromosomes (25 total):
chr1~chr22,chrX,chrY,chrM(mitochondrial) - Mitochondrial gene prefix: Gene symbols usually start with
MT- - Ribosomal gene prefixes: Ribosomal protein genes usually start with
RPL,RPS,MRPL,MRPS - Immune-related gene prefixes: B-cell receptor genes mainly start with
IGH,IGK,IGL; T-cell receptor genes mainly start withTRA,TRB,TRG,TRD - Other common QC-related prefixes: Hemoglobin genes, which may indicate red blood cell contamination, mainly start with
HBA,HBB; heat shock protein genes, which may indicate cellular stress, mainly start withHSP
The current reference genome has undergone customized preprocessing by the SeekGene platform. Most inactive background pseudogenes have been proactively removed to greatly reduce noise. After processing, a total of 44,528 genes are retained:
| Category | Gene Type | Gene Count |
|---|---|---|
| Protein-coding | protein_coding | 20,046 |
| Non-coding RNA | lncRNA | 18,866 |
Small RNAs such as misc_RNA, snRNA, miRNA, snoRNA | 7,112 | |
| Immunoglobulin (B-cell receptor) | IG_V_gene, IG_D_gene, IG_J_gene, IG_C_gene, and corresponding pseudogenes | 409 |
| T-cell receptor | TR_V_gene, TR_J_gene, TR_C_gene, TR_D_gene, and corresponding pseudogenes | 239 |
1.2 Mouse
- Genome assembly: GRCm39
- Annotation version: GENCODE vM38
- Original download link: GENCODE vM38 GTF
Chromosome and gene statistics
- Chromosomes (22 total):
chr1~chr19,chrX,chrY,chrM(mitochondrial) - Mitochondrial gene prefix: Gene symbols usually start with
mt- - Ribosomal gene prefixes: Ribosomal protein genes usually start with
Rpl,Rps,Mrpl,Mrps - Immune-related gene prefixes: B-cell receptor genes mainly start with
Igh,Igk,Igl; T-cell receptor genes mainly start withTra,Trb,Trg,Trd - Other common QC-related prefixes: Hemoglobin genes mainly start with
Hba,Hbb; heat shock protein genes mainly start withHsp; in addition, many predicted genes start withGm(Gene Model), and many RIKEN cDNA clones end withRik
Consistent with the human reference genome processing strategy, after removal of background pseudogenes, 61,449 genes are retained in the mouse reference genome:
| Category | Gene Type | Gene Count |
|---|---|---|
| Protein-coding | protein_coding | 21,760 |
| Non-coding RNA | lncRNA | 32,889 |
Small RNAs such as miRNA, snoRNA, snRNA, misc_RNA | 5,728 | |
| Immunoglobulin (B-cell receptor) | IG_V_gene, IG_D_gene, IG_J_gene, IG_C_gene (including LV and pseudogenes) | 426 |
| T-cell receptor | TR_V_gene, TR_J_gene, TR_C_gene, TR_D_gene, and corresponding pseudogenes | 270 |
1.3 Rat
- Genome assembly: GRCr8
- Annotation version: Ensembl release 116
- Original download link: Ensembl 116 GTF
Chromosome and gene statistics
- Chromosomes (26 total):
chr1~chr20,chrX,chrY,chrMT(mitochondrial) - Unlocalized scaffolds:
chrJAXUCZ010000023.1,chrJAXUCZ010000028.1,chrJAXUCZ010000075.1 - Mitochondrial gene prefix: Gene symbols usually start with
MT- - Ribosomal gene prefixes: Ribosomal protein genes usually start with
Rpl,Rps,Mrpl,Mrps - Other common QC-related prefixes: Hemoglobin genes mainly start with
Hba,Hbb; heat shock protein genes mainly start withHsp; in addition, many uncharacterized genes start withLOC
Because this rat annotation file has already undergone specific filtering, only 1 core gene type is currently retained by default:
| Category | Gene Type | Gene Count |
|---|---|---|
| Protein-coding | protein_coding | 43,360 |
2. General Reference Genome Download Rules from International Databases
To help users obtain reference genomes for other species or confirm file versioning rules before analysis, we summarized the standard download formats used by two major databases: NCBI RefSeq and Ensembl.
2.1 NCBI RefSeq
NCBI uses a three-level numeric directory structure (for example, 002/007/445/) to improve FTP server retrieval performance. This segmented directory is derived from the full accession number (for example, GCF_002007445).
Download file formats:
- FASTA file (
.fna.gz)https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/{Accession segments}/{Full accession}/{Full accession}_genomic.fna.gz - GTF file (
.gtf.gz)https://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/{Accession segments}/{Full accession}/{Full accession}_genomic.gtf.gz
URL structure example (giant panda, Ailuropoda melanoleuca):
| Component | Example | Description |
|---|---|---|
| Base path | https://ftp.ncbi.nlm.nih.gov/genomes/all/ | NCBI genome download server |
| Segmented directory | GCF/002/007/445/ | Accession number split into groups of three digits |
| Full accession | GCF_002007445.1_ASM200744v2 | GCF number + version (.1) + assembly name |
| File suffix | _genomic.fna.gz / _genomic.gtf.gz | FASTA sequence file or GTF annotation file |
2.2 Ensembl
Ensembl organizes data by release version. Species names are written in lowercase Latin binomials, with spaces replaced by underscores.
Download file formats:
- FASTA file (
.fa.gz)https://ftp.ensembl.org/pub/release-{version}/fasta/{lowercase_species_name}/dna/{Species.Assembly.dna.toplevel.fa.gz} - GTF file (
.gtf.gz)https://ftp.ensembl.org/pub/release-{version}/gtf/{lowercase_species_name}/{Species.Assembly.{version}.gtf.gz}
URL structure example (Vombatus ursinus):
| Component | Example | Description |
|---|---|---|
| Base path | https://ftp.ensembl.org/pub/release-115/ | Ensembl FTP server and release version |
| Data type | fasta/ or gtf/ | Data type directory |
| Species directory | vombatus_ursinus/ | Species name in lowercase with spaces replaced by underscores |
| Subdirectory | dna/ (FASTA only) | Fixed subdirectory for FASTA sequence files |
| Filename | Vombatus_ursinus.bare-nosed_wombat_genome_assembly.dna.toplevel.fa.gz | Capitalized species name + assembly name |
Note: Assembly names vary across species and may follow different naming styles, such as
bare-nosed_wombat_genome_assemblyorVulVul2.2. Please use the actual naming shown in Ensembl for the target species when constructing the filename.
3. SeekGene Cloud Reference Genome Resources
To facilitate use, the standard reference genome files listed above, as well as many other commonly used files, have been uploaded to the official SeekGene OSS repository. They can be directly called on the cloud platform or referenced in customized analysis workflows.
To make direct downloading easier for other users, the currently built and publicly available reference genome packages are listed below:
| Type | Species | Filename | Download Link |
|---|---|---|---|
| ARC | Human | refdata-arc-GRCh38-2024-A-v44.zip | Download |
| ARC | Mouse | refdata-arc-GRCm39-2025-A-vM38.zip | Download |
| ARC | Rat | refdata-arc-GRCr8_Ensembl_116.zip | Download |
| GEX | Human | refdata-gex-GRCh38-2024-A-v44.zip | Download |
| GEX | Mouse | refdata-gex-GRCm39-2025-A-vM38.zip | Download |
| GEX | Rat | refdata-gex-GRCr8_Ensembl_116.zip | Download |
| MET | Human | refdata-met-GRCh38-2024-A-v44.zip | Download |
| MET | Mouse | refdata-met-GRCm39-2025-A-vM38.zip | Download |
| MET | Rat | refdata-met-GRCr8_Ensembl_116.zip | Download |
| GEX + VDJ | Human | refdata-gex-vdj-GRCh38-2024-A-v44_multi.zip | Download |
| GEX + VDJ | Mouse | refdata-gex-vdj-GRCm39-2025-A-vM38_multi.zip | Download |
Summary
When processing standard transcriptome reference annotations, the SeekGene platform applies a strict customized filtering strategy to maximize the signal-to-noise ratio of single-cell and spatial sequencing data. For example, purely background pseudogenes such as processed_pseudogene and unprocessed_pseudogene are removed, while biologically meaningful gene types are retained.
These retained genes mainly include protein-coding genes (protein_coding), various classes of long and small non-coding RNAs (lncRNA, miRNA, etc.), and all gene segments involved in T/B cell receptor V(D)J recombination, including their corresponding pseudogenes such as IG_V_pseudogene, because they still provide important reference value for immune repertoire assembly.