Genomics Expert
Before Starting
- What organism and genome size?
- Short-read or long-read sequencing?
- Whole genome, exome, or targeted?
Core Expertise Areas
Sequencing Technologies
Short-read: Illumina — high accuracy, low cost, 150-300bp reads. Long-read: PacBio (SMRT), Oxford Nanopore — spans repeats, structural variants. Hi-C: chromatin conformation, scaffolding assemblies. Single-cell: 10x Genomics, Drop-seq — individual cell transcriptomes.
Genome Assembly
De novo assembly: overlap-layout-consensus, de Bruijn graphs. Tools: SPAdes (short-read), Flye (long-read), Hifiasm (HiFi). Assembly metrics: N50, L50, contig count, genome completeness (BUSCO). Scaffolding: Hi-C, optical mapping, genetic maps.
Genome Annotation
Structural annotation: gene prediction (Augustus, MAKER), repeat masking (RepeatMasker). Functional annotation: BLAST, InterPro, GO terms, KEGG pathways. Non-coding elements: promoters, enhancers, regulatory regions. Comparative annotation: synteny, ortholog identification.
Variant Calling
SNVs and indels: GATK HaplotypeCaller, DeepVariant. Structural variants: Manta, LUMPY, PBSV. Copy number variants: CNVkit, Control-FREEC. Variant filtering: VQSR, hard filters, population frequency.
Functional Genomics
RNA-seq: transcript quantification, differential expression (DESeq2, edgeR). ChIP-seq: protein-DNA binding, peak calling (MACS2). ATAC-seq: chromatin accessibility, open chromatin regions. Metagenomics: community profiling, functional annotation of environmental samples.
Key Patterns
Best Practices
- Always check sequencing QC before assembly or mapping
- Use BUSCO to assess assembly and annotation completeness
- Filter variants by quality score and depth before analysis
- Validate structural variants with orthogonal evidence
- Document software versions for reproducibility
Common Pitfalls
| Pitfall | Fix |
|---|---|
| Low coverage causing false variants | Aim for 30x minimum for WGS |
| Repeat regions causing assembly gaps | Use long reads to span repeats |
| Contamination in assemblies | Run BLAST against contaminant databases |
| Ignoring batch effects in RNA-seq | Include batch as covariate in model |
Related Skills
- molecular-biology-expert
- bioinformatics-expert
- genetics-expert
- computational-chemistry-expert