I’m an MSc Genomics graduate interested in variant interpretation, computational genomics, transcriptomics and reproducible bioinformatics.
My projects usually begin with a biological question and follow it through data preparation, statistical analysis, visualisation and interpretation. I mainly work in R, Python and Bash, and I try to make the reasoning behind an analysis as clear as the final output.
My current focus is pathogenicity prediction and the use of sequence and structural information to understand how genetic variants may affect biological function.
A reconstruction of my MSc dissertation analysis, which assessed whether AlphaRING v2 could distinguish SARS-CoV-2 spike variants associated with decreased viral fitness from variants with neutral or increased fitness. The final curated analysis produced an area under the receiver operating characteristic curve of approximately 0.84, while SHAP values indicated that predicted stability change and residue interaction degree were the strongest contributors to model predictions.
A count-level RNA-seq analysis comparing dexamethasone-treated and untreated human airway smooth muscle cells. After low-count filtering, DESeq2 tested 22,369 genes and identified 2,694 as significantly differentially expressed at an adjusted p-value below 0.05.
An analysis of expression variation and immune-related transcriptional activity across 1,097 TCGA-BRCA tumour samples. The workflow combines principal component analysis with a nine-gene immune signature and separates the cohort into relative immune-high and immune-low groups for exploratory comparison.
A command-line workflow that takes paired-end Escherichia coli reads through quality control, alignment, BAM processing, variant calling and filtering. The analysis produced 792 raw candidate variants, of which 407 remained after filtering.
A workflow-managed version of the bacterial variant-calling analysis. I used Nextflow to separate each analytical stage, manage dependencies between processes and retain pipeline reports alongside the final VCF summaries.
A set of lightweight Python command-line tools for inspecting FASTA, FASTQ and VCF files. The programs calculate sequence statistics, summarise read quality, count candidate variants and generate reverse-complement sequences using real inputs from the bacterial variant-calling project.
- Languages: R, Python and Bash
- Genomic data: FASTA, FASTQ, BAM and VCF processing
- Transcriptomics: DESeq2, differential expression analysis and gene-expression visualisation
- Variant analysis: variant calling, pathogenicity prediction and structural interpretation
- Statistics: principal component analysis, receiver operating characteristic analysis and multiple-testing correction
- Reproducibility: Conda environments, Nextflow, ordered analysis scripts and documented outputs