Grc Human Genetics And Genomics

Genome Reference Consortium (GRC) resources provide the standardized human reference genome assemblies used in nearly all modern human genetics and genomics research. This guide explains how GRC reference genomes function in variant discovery, polygenic risk scoring, and clinical interpretation. It is intended for bioinformatics trainees, clinical geneticists, and laboratory scientists who need a practical, source bounded framework for using GRC based data.

The foundational reference genome from the GRC is used to align short and long read sequencing data, call variants, and assign clinical significance. According to NCBI Bookshelf, the reference genome is a composite sequence from several donors, not a single individual, which introduces important interpretive constraints. Official training from EMBL EBI emphasizes that all downstream analyses from RNA expression quantification to copy number detection depend on the accuracy of that initial alignment against a GRC assembly.

At a Glance

Core Concepts of GRC in Human Genetics

The GRC releases updated assemblies to correct sequence errors and fill gaps. The current major version is GRCh38 (also called hg38), which replaced GRCh37 (hg19) in 2013. Newer patches are released regularly. When you download a reference genome from an official repository, you must also include alternative allele sequences and decoy contigs. The Galaxy Training Network stresses that omitting these sequences leads to false positive variant calls in regions like the HLA complex.

Each GRC assembly includes a primary chromosome set, unlocalized scaffolds, and unplaced contigs. The primary set is the most commonly used for short read alignment. However, for clinical applications involving known medically relevant genes, using the full assembly with alt loci improves accuracy. The Bioconductor package BSgenome.Hsapiens.NCBI.GRCh38 provides the complete reference and is updated with each GRC patch.

A practical understanding of GRC resources also requires knowledge of the reference genome's limitations. The genome is a haploid representation, but humans are diploid. This means that heterozygous sites can only be discovered by comparing reads to the single reference, not by direct alignment to the alternative allele. Researchers who analyze polygenic scores for conditions like prostate cancer aggressiveness rely on GRC based variant calls from large cohorts as described in Germline polygenic score for prostate cancer aggressiveness. The reference genome determines which variants are included in the scoring model.

Decision Points for Reference Genome Assembly

Selecting which GRC assembly to use is a critical decision that affects reproducibility and cross study comparability.

First, decide between GRCh37 and GRCh38. GRCh37 is still used by many legacy databases and clinical pipelines. If your project requires comparing against historical GWAS summary statistics that were mapped to GRCh37, you may need to lift over coordinates. However, GRCh38 offers fewer errors, better representation of segmental duplications, and inclusion of centromeric sequences. The NCBI Sequence Read Archive stores both alignment types, so check the metadata of your dataset before starting.

Second, decide whether to include alt and decoy sequences. For most RNA seq analyses, the primary assembly is sufficient. For DNA variant calling in highly polymorphic regions such as the MHC locus, include alt sequences to avoid false positive calls. The EMBL EBI Training modules on variant calling show that omitting the decoy sequence causes reads from novel insertions to map incorrectly to other genomic regions, inflating coverage and generating spurious signals.

Third, for structural variant detection, consider using a graph genome approach instead of a linear reference. The GRC has released a draft pangenome, but linear references remain the standard in most clinical labs. For example, a study on prenatal duplication interpretation used GRCh37 for family based analysis as described in Family Based Interpretation of a Prenatally Detected 15q11.2 Duplication. The authors explicitly state that the duplicated region boundaries were defined using the UCSC Genome Browser aligned to GRCh37.

Practical Workflow: Aligning Reads to a GRC Reference

Below is a step by step implementation sequence for aligning whole genome sequencing reads to a GRC reference.

1. Download the correct reference. Obtain the FASTA file for GRCh38 from the NCBI FTP site or use a Bioconductor package. Index the reference with BWA or Bowtie2. For long reads, use Minimap2 with the appropriate preset.
2. Preprocess raw reads. Trim adapters and low quality bases using tools like Fastp or Trimmomatic. The quality of the input directly affects alignment accuracy. Galaxy training materials recommend removing bases with quality below Q20.
3. Align reads to the reference. Run the aligner with parameters that match your sequencing technology. For the Burrows Wheeler Aligner (BWA MEM), use default settings for paired end reads. For splice aware alignment in RNA seq, use STAR or HISAT2 with the GRC annotation GTF file.
4. Post process the alignment. Sort the BAM file by coordinate, mark duplicates using Picard or sambamba, and index the file. Compute alignment statistics such as total reads mapped, percentage properly paired, and median insert size.
5. Variant calling. Use GATK HaplotypeCaller or freebayes following best practices. These tools use the reference to define candidate variants. For joint genotyping, combine GVCFs across samples.

The Galaxy Training Network provides interactive workflows for each step, including a ready to use workflow for human germline variant calling from whole genome sequencing. Following these steps with the correct GRC reference ensures that your variant calls are compatible with public resources like the 1000 Genomes Project and ClinVar.

Quality Checks and Validation

After alignment and variant calling, quality checks must be performed to ensure that results are reliable for downstream interpretation.

Alignment level quality. Check the percentage of mapped reads. For high quality whole genome sequencing, expect more than 95% mapped. Inspect the coverage uniformity using the genome coverage tool in BEDTools. Regions with zero coverage may indicate mapping problems or genuine gaps in the reference. The Bioconductor package Rsamtools can generate per base coverage statistics.

Variant level quality. Use the variant quality score recalibration (VQSR) approach from GATK to filter false positives. Examine the transition to transversion ratio (Ti/Tv). For whole genome data, the Ti/Tv ratio should be around 2.0 to 2.1. Ratios significantly lower suggest many false positive calls. For exome data, the ratio is higher, around 3.0.

Clinical validation. If your analysis is for clinical interpretation, orthogonal validation using Sanger sequencing or a different sequencing technology is recommended. In a study of laminopathies, researchers combined GRC based variant calls with clinical risk prediction models as shown in Laminopathies: natural history and risk prediction of heart failure. They validated a subset of variants by Sanger sequencing before using them in their model.

Check for reference bias. Compare alignment results between GRCh37 and GRCh38 for a small region to confirm that the newer assembly does not introduce unexpected changes in variant annotation. This is especially important for structural variant analysis.

Common Mistakes in Reference Based Analysis

Using the wrong reference version. Many public datasets are aligned to GRCh37. If you align your data to GRCh38 and compare directly without liftover, you will get false negative or false positive results. Always verify the reference version used in the original publication.

Ignoring reference patches. GRC releases patches (e.g., GRCh38.p14). Using an older patch can mean missing corrected bases. For instance, the GRCh38.p14 patch corrected several errors in the Y chromosome. Laboratories that do not update their reference will miscall variants in those regions.

Assuming the reference is "normal". The GRC reference is a composite and contains rare alleles. A variant that is listed as "reference" may not be the major allele in your population. Researchers who studied ATOH1 variants linked to neurodevelopmental disorders and hearing loss explicitly compared their findings against the reference population in gnomAD, not against the GRC sequence itself, as reported in Dominant and recessive ATOH1 variants cause distinct neurodevelopmental disorders with hearing loss. Always compare allele frequencies against population specific databases.

Using a default reference for all applications. For example, mitochondrial DNA analysis requires a separate circular reference (rCRS). Using the nuclear reference will lead to misalignment. Similarly, for long read sequencing of repetitive regions, a linear reference may cause high mismatch rates. Graph genome approaches may be more appropriate.

Limits and Uncertainty in Interpretation

GRC reference genomes are incomplete. Over 200 known gaps remain in GRCh38, particularly at centromeres and acrocentric short arms. The T2T consortium has closed these gaps in the CHM13 reference, but GRC resources are still the official human reference for most databases and clinical guidelines.

Reference bias for structural variation. Large insertions present in your sample but absent from the reference will not be detected by alignment based methods. This is a fundamental limitation. Methods like joint calling across multiple samples can partially mitigate this, but the variant will be represented as a deletion relative to the reference.

Uncertainty in clinical classification. A variant that appears as a mismatch against the reference may be a sequencing error, alignment artifact, or true variant. Population based allele frequency and functional prediction tools must be used. The EMBL EBI Training course on variant interpretation emphasizes using multiple lines of evidence and caution when the variant lies in a repetitive region.

Polygenic scores and ancestry. GRC based variant calls can be used to compute polygenic scores, but the score performance varies across ancestries. The prostate cancer polygenic score reported in Germline polygenic score for prostate cancer aggressiveness was derived from a predominantly European cohort and its accuracy in African populations is uncertain. Reference bias and differing linkage disequilibrium patterns contribute to this limitation.

Allelic dropout in targeted sequencing. PCR primers for amplicon based methods are designed against the reference sequence. If a sample has a deletion in the primer binding region, the allele will not be amplified, leading to false homozygosity. This is a well known pitfall that cannot be corrected by alignment alone.

Frequently Asked Questions

Q: What does GRC stand for in human genetics? A: GRC stands for Genome Reference Consortium, an international group that maintains and updates the human reference genome assembly used for sequence alignment and variant calling.

Q: Should I use GRCh37 or GRCh38 for my project? A: For new projects, use GRCh38 because it has fewer errors and better representation of complex regions. Only use GRCh37 if you need to compare directly with legacy datasets that cannot be lifted over reliably.

Q: Can I use the GRC reference for RNA seq analysis? A: Yes, but you must also use the corresponding gene annotation GTF file from GENCODE or RefSeq. The alignment software will use the reference sequences and the annotation to map reads that span exon intron junctions.

Q: How often should I update my local reference copy? A: Update when a new major or minor GRC patch is released, especially if you are analyzing clinical samples. Check the NCBI Announcements page quarterly. Using an outdated reference may cause you to miss known sequence corrections.

References and Further Reading

• NCBI Bookshelf on the Human Genome Reference
• EMBL EBI Training: Reference Genomes
• Galaxy Training Network: Reference Genome Preparation
• Bioconductor BSgenome.Hsapiens.NCBI.GRCh38
• NCBI Sequence Read Archive: GRC reference usage
• Germline polygenic score for prostate cancer aggressiveness (medRxiv)
• Family Based Interpretation of a Prenatally Detected 15q11.2 Duplication (Cureus)
• Laminopathies: natural history and risk prediction of heart failure (Eur Heart J)
• Dominant and recessive ATOH1 variants (Am J Hum Genet)

Related Articles

• Protein Synthesis
• Incomplete Dominance Gene
• Cell Membrane Function Biology
• Dna Structure
• Protein Structure