Guided case study
Module 1 · Understanding omics · Self-checked
One brain tumour, measured in every layer you have studied. Work through the six tasks: write your own answer first, then open the model answer and tick the task off.
This case brings together everything in Module 1. The patient is invented for teaching, but every result in the case file is the kind a real brain tumour team would see, and the biology behind them is well established. Read the whole case file once before you start the tasks.
- Patient
- 34 years old; first seizure; MRI shows a mass in the left frontal lobe. Surgery removes most of it.
- Pathology
- A diffuse glioma. IHC for the mutant protein IDH1 R132H is positive; ATRX protein is lost.
- Panel sequencing
- IDH1 c.395G>A (p.R132H): 228 of 600 reads carry the variant. Tumour purity estimated at 80%. A TP53 mutation is also present.
- Copy number
- No combined loss of chromosome arms 1p and 19q (“1p/19q intact”).
- Methylation array
- Widespread CpG island hypermethylation. The classifier matches “astrocytoma, IDH-mutant”. MGMT promoter methylated.
- MR spectroscopy
- A peak consistent with 2-hydroxyglutarate (2-HG) in the tumour.
Terms in the case file
- Glioma is a tumour that arises from glial cells, the supporting cells of the brain. Diffuse means that the tumour cells spread into the surrounding brain instead of forming a sharp edge, which is why surgery can rarely remove all of it.
- MRI (magnetic resonance imaging) makes detailed pictures of soft tissue such as the brain. MR spectroscopy uses the same kind of scanner to measure the chemical make-up of a chosen region, as a set of peaks, each belonging to a molecule. It works on the same principle as NMR (Proteomics and metabolomics), in a living patient.
- c.395G>A (p.R132H) is the standard way of writing a variant. The c. part describes the DNA change: at position 395 of the gene's coding sequence, G is replaced by A. The p. part describes the protein change: amino acid 132, arginine (R), becomes histidine (H).
- ATRX and TP53 are genes; "ATRX protein is lost" means the IHC stain found no ATRX protein in the tumour cells, a sign that the gene is not working.
- 1p and 19q are chromosome arms. Each chromosome has a short arm, called p, and a long arm, called q. "Combined loss" means both the short arm of chromosome 1 and the long arm of chromosome 19 are missing.
- Hypermethylation means more methylation than normal. MGMT is a gene whose promoter can be methylated, like MLH1 in Epigenomics.
- Tumour purity is the share of cells in the sample that are tumour cells; the rest are normal brain, blood vessels and immune cells. A purity of 80% means about one cell in five is normal tissue.
- A heterozygous mutation is present in one of the two copies of a gene in a cell. A mutation is clonal if every tumour cell carries it and subclonal if only some do. You met these ideas with variant allele fraction in Genomics.
Go deeper: the grade, and the mutation the stain can miss
Under the 2021 WHO classification of brain tumours, an IDH-mutant astrocytoma is given a grade of 2, 3 or 4. Grade mostly follows how the tissue looks, with one molecular exception: if both copies of CDKN2A and its neighbour CDKN2B are deleted, the tumour is grade 4 whatever the microscope shows. CDKN2A is the brake on cell division you met in Epigenomics. The same edition also reserves the name glioblastoma for IDH-wildtype tumours only.
The IHC stain in this case recognises one exact change, R132H. That change accounts for roughly nine in ten IDH1 mutations in gliomas, but not all of them. Rarer changes at the same position, such as R132C, are invisible to the antibody. That is why a negative stain in a younger adult with a diffuse glioma is usually followed by sequencing of IDH1 and IDH2.
How to approach this
Go through the case file one line at a time and, for each line, ask the question from The central dogma, as data: which molecule did this test measure? Then follow the links between layers in the direction information flows: from the DNA change, to the enzyme it alters, to the metabolite that enzyme makes, and on to the epigenome. Write a short answer for each task before you open the model answer, and use the numbers in the case file; tasks 2 and 3 can be checked by arithmetic and by reasoning from earlier lessons. If you get stuck, the lesson numbers under each task title say where to look.
Why work in this order? The case file lists results by test, the way a hospital report arrives, but the biology runs the other way, from a cause in the DNA to its effects. Reading each test in isolation gives you a list of facts; following the flow turns them into an explanation, and shows you which results support each other.
Two habits help on real cases. First, note which layer each conclusion rests on, so you can tell a measured result from an inferred one. Second, look for results that disagree. A mismatch, such as a mutation the stain did not show, is often where the most useful question lies, and it is easy to skip over when most results agree.
Map the layers
The central dogma, as data and Proteomics and metabolomics
For each result in the case file, name the layer it measures and the assay used.
Hint. Use the table in The central dogma, as data. For each test, decide whether it reads DNA sequence, marks on DNA, RNA, protein or small molecules. Remember from Proteomics and metabolomics that IHC uses antibodies, and check whether every layer in the table is covered by at least one test.
Check your answer
IHC: proteome, by antibody on tissue. Panel sequencing and copy number: genome. Methylation array: epigenome. MR spectroscopy: metabolome, measured in the living patient. No transcriptome assay was done; that layer is inferred.
Note that one technology can serve more than one layer. The same scanner gave the MRI picture of the mass and the chemical spectrum, and sequencing reads gave both the mutations and, in many labs, the copy-number result. Name the layer by what was measured, not by the machine.
Read the VAF
Genomics
Calculate the VAF for IDH1 R132H. Compare it with what you expect for a heterozygous mutation in every tumour cell at 80% purity. Is it clonal?
Hint. Follow the worked example in Genomics exactly, with this case's numbers: divide the variant reads by the total reads, then work out purity ÷ 2. If the two values are close, think about what that says about how many tumour cells carry the mutation, and so about when it arose.
Check your answer
VAF = 228 ÷ 600 = 0.38. Expected = 0.80 ÷ 2 = 0.40. They are close, so the mutation is in essentially all tumour cells: clonal, and probably the earliest event. That fits what is known about IDH-mutant gliomas.
A different result would have meant something different. A VAF of about 0.20 at the same purity would suggest that only around half the tumour cells carry the mutation, a subclonal, later event. A VAF well above 0.40 could mean that the normal copy of the gene had been lost. Comparing the observed and expected values is what lets you tell these apart.
Genome to metabolome
Proteomics and metabolomics
The IDH1 mutation changes an enzyme. Explain how that produces the MR spectroscopy result.
Hint. Reread the IDH bullet in the metabolomics section of Proteomics and metabolomics. An enzyme turns one small molecule into another. Ask what the normal enzyme makes, what the mutant enzyme makes instead, and which line of the case file detects that product.
Check your answer
Normal IDH1 makes α-ketoglutarate. The R132H mutant gains a new activity: it converts α-ketoglutarate into 2-hydroxyglutarate, which builds up to very high levels. MR spectroscopy can detect it without surgery.
Metabolome to epigenome
Epigenomics
Why might this tumour show widespread CpG island hypermethylation?
Hint. Proteomics and metabolomics says the new metabolite "changes the epigenome", and Epigenomics explains that methylation marks are added and removed by enzymes. Consider what would happen to the overall level of methylation if the enzymes that remove it were held back, and how the metabolite from task 3 might be involved.
Check your answer
2-HG blocks enzymes that depend on α-ketoglutarate, including the TET enzymes that remove DNA methylation and some histone demethylases. Methylation builds up across CpG islands, silencing genes. A single DNA change has reshaped the epigenome through a metabolite.
Classify the tumour
Genomics and Epigenomics
The WHO classification of brain tumours uses molecular results. Which findings decide between astrocytoma and oligodendroglioma here?
Hint. The WHO classification is the international standard, published by the World Health Organization, for naming tumours. Start by listing what the two tumour types share, then look for the copy-number result in the case file and the genes whose loss or mutation is reported. Epigenomics explains why the methylation classifier can act as an independent check.
Check your answer
Both are IDH-mutant. Oligodendroglioma also has 1p/19q codeletion; this tumour is 1p/19q intact, with ATRX loss and a TP53 mutation, which fit astrocytoma. The methylation classifier agrees: several layers point the same way.
Which results matter for treatment?
Epigenomics and Proteomics and metabolomics
Name two results that could influence treatment, and explain why.
Hint. Look for one result that names a changed protein a drug could act on directly, and one epigenetic result that silences a gene, which Epigenomics showed can matter as much as a mutation. For the second, ask what the silenced gene normally does, and how losing it might change the effect of a treatment that damages DNA.
Check your answer
The IDH mutation itself: vorasidenib, an inhibitor of mutant IDH1 and IDH2, was approved in 2024 for some IDH-mutant gliomas. MGMT promoter methylation: it silences a DNA repair gene and is associated with better response to temozolomide chemotherapy. Treatment is decided by the clinical team; the point is that each layer added information.
Notice that the two results come from different layers and were found by different tests. The IDH mutation was seen by both the stain and the panel, while MGMT methylation appears only on the methylation array. A team that had run DNA sequencing alone would have missed it, the point made in the key message of Epigenomics.
What the case shows
Look back at the six tasks together. One DNA change, seen in the genome, explained a result in the metabolome and a pattern in the epigenome, and each layer was measured with a different technology. No single test told the whole story: the IHC stain, the panel, the copy-number result and the methylation classifier each settled part of the diagnosis, and they agreed. That agreement across layers is what makes the conclusion trustworthy. Notice also what was missing: no RNA was measured, so anything about gene activity here is inferred, not observed.
An RNA measurement would have been the natural next test. Bulk RNA-seq, as in Bulk transcriptomics, could show whether the genes silenced by hypermethylation are really quiet in this tumour. Single-cell and spatial biology could go further and show which cells in the mix carry the tumour programme and where the normal and immune cells sit. Each added layer can confirm or challenge what the others suggest.
Module 2, Biological data and databases, builds on this case. It covers the files and databases, and Module 3 then teaches the command line, used to handle those files at scale. Every result here would, in real work, be stored in a file and looked up in public resources. You will meet the file formats that hold sequences, annotations and variants, the main databases for genes, proteins and structures, and the identifiers that link one gene across all of them. You will also learn to read a gene record, a protein record and a published paper critically, so that a variant such as IDH1 R132H becomes something you can trace for yourself.
Take it further
Look up IDH1 in cBioPortal and see which cancer types carry the R132H mutation most often. Then learn to analyse omics data yourself in Module 5: sequencing data.
When you look IDH1 up, compare the numbers for gliomas with those for other cancer types, and notice how often the same single change, R132H, accounts for the mutations seen. Recurrence like this at one exact position is the pattern of a driver described in Genomics.
Further reading · Module 1Sources for this module and the rest of the course are gathered on one page. · All reading for this course
Up next
Lesson 1
File formats you will meet: FASTA, GenBank, GFF, VCF, PDB