Day 10 discussion seeds - mestato/EPP622_2026 GitHub Wiki

What is the difference between SNP calling and genotype calling? (IE what is a SNP vs a genotype?)

You are a SNP and genotype caller. Is there a SNP and what is your genotype call?

• Easy:
• Scenario 1. The reference genome has an A. 10 sample reads all have a T.
• Scenario 2. The reference genome has an A. 19 sample reads have a T, 1 sample read has an G. The quality of the G base is phred 10.
• Scenario 3. The reference genome has an A. 7 sample reads have a T, 3 sample reads have an A. The quality of the A and T bases are phred 30+.
• More Difficult:
• Scenario 3. The reference genome has an A. 15 sample reads have a A, 2 sample reads have an T. The phred of Ts is 10 and 20.
• Scenario 4. The reference genome has an A. 2 sample reads have an A, 1 sample read has a T. The T base has a phred of 20. No prior information available.
• Scenario 5. The reference genome has an A. 2 sample reads have an A, 1 sample read has a T. The T base has a phred of 20. This location is known to have a SNP with alleles of A or T.
• Scenario 6. The reference genome has an A. 2 sample reads have an T with a phred of 30.

Thinking about scenario 4 vs 5, what is the benefit of doing snp calling prior to genotype calling? (Assuming you have more than one sample and/or a database of known SNPs.)

These are the main causes of errors in SNP calling. Why and how can you fix/avoid?

• incorrect base call in read
• incorrect alignment of read to reference - read in wrong location
• incorrect alignment of read to reference - local alignment problem (we'll read more about this soon)
• incorrect base in reference genome (assembly error)
• incorrect structure in assembly (collapsed repeats)
• structural variation in sequenced sample
• lack of sequencing depth in sequenced sample
• duplicate reads (i.e. - too little DNA going in, leading to major PCR artifacts)

Compare your fastqc results from Tuesday. Do they match? Are you concerned about any samples?