Which genes belong in the background set?

• Every gene in the genome
• Only genes that stood a chance of selection
• Only the changed genes

What does an over-representation test ask about your gene list?

• Whether a pathway holds more changed genes than chance predicts
• Whether each single gene is truly changed
• Whether the assay measured every gene

Long transcripts tend to attract more reads, so they are called changed more often.

Circle one:   True   False

A naive test flags mostly long, highly expressed genes. What should you do next?

• Model each gene's prior chance from length and expression, then retest
• Drop the long genes and keep the result
• Report the terms as biology

Your assay measured 12,000 genes but you filtered out 2,000 before testing. What is the right background?

• Only the changed genes
• All genes in the genome
• The 10,000 measured and kept genes

How does a rank-based method differ from an over-representation test?

• It tests a hard cutoff shortlist
• It walks the full ordered list and spots pathway members near the top
• It skips the background set entirely

Using the whole genome as background is fine even when your assay measured only half the genes.

Circle one:   True   False

A term is significant under one permissive setting but vanishes with a proper background and a rank-based check. What do you conclude?

• Average the two results and report both
• Follow it up at once
• Treat it as a methods artefact, not a finding