Genomics, proteomics, and data-driven biology
Beyond single genes, bioinformatics studies whole genomes, all proteins, and the networks linking them.
Genomics
Genomics studies entire genomes. Comparing genomes shows which genes species share and which are unique. Genome-wide association studies (GWAS) scan the genomes of many people to find genetic variants more common in those with a disease, highlighting genes worth studying.
Transcriptomics
Transcriptomics measures which genes are being expressed, by sequencing all the RNA in a sample (RNA-seq). Comparing tumour and healthy tissue shows which genes are switched on or off in cancer. Single-cell RNA-seq reveals differences between individual cells in a tissue.
Proteomics
Proteomics identifies and measures the proteins in a sample, often using mass spectrometry, which weighs fragments of proteins very precisely. Because proteins are modified after translation and vary in amount, the proteome reveals cell activity more directly than the genome alone.
Protein structure prediction
Predicting a protein's 3D shape from its sequence was a challenge for 50 years. In 2020, the AI system AlphaFold predicted structures with accuracy close to experiments, and its creators shared predictions for over 200 million proteins. Its developers shared the 2024 Nobel Prize in Chemistry.
Systems biology
Systems biology studies how genes, proteins, and metabolites interact in networks, rather than one at a time. Network maps show hub proteins connected to many others; disrupting hubs is often more harmful. Models can predict how a cell responds to a drug or mutation.
Personalised medicine
Genome information can guide treatment. Some cancer drugs target specific mutations found by sequencing the tumour. Pharmacogenomics studies how genetic variants affect drug responses, for example in how fast people break down certain medicines, so doses can be tailored.
Machine learning in biology
Machine learning finds patterns in large datasets: predicting gene function, classifying cell types, spotting disease in medical images, and designing new proteins. Results must be validated experimentally, and biased or limited training data can give misleading predictions.
Data ethics
Genetic data are personal and can reveal information about relatives. Concerns include consent, privacy, discrimination by employers or insurers, and fair representation: most genomic studies have focused on people of European ancestry, so findings may apply less well to others. Laws and data-sharing policies try to balance research benefits and privacy.
Key takeaways
Genomics compares whole genomes, and GWAS links variants to disease. Transcriptomics measures gene expression; proteomics measures proteins by mass spectrometry. AlphaFold transformed structure prediction. Systems biology studies networks. Genomic data enable personalised medicine, aided by machine learning, but raise privacy and fairness concerns.
Flashcard
Study of whole genomes
Check your understanding
Why might a genetic risk prediction be less accurate for someone of non-European ancestry?