de novo assembly

Question 1

Explain the greedy approach

and drawbacks

Answer 1

Pairwise alignment
Find the ones that have overlap and merge
Repeats only found if small, high computational cost

Question 2

Explain approach of Overlap Layout Consensus and de Bruijn

Answer 2

Correct sequencing errors
Assemble contigs
Combine contigs to scaffolds

Question 3

OLC, method, drawbacks

Answer 3

Make graph where nodes are reads
Branch graph if there are overlaps between reads
Not good for short reads and repeats

Question 4

de Bruijn, method and drawbacks

Answer 4

Map kmers instead of entire reads. Every kmer exists once, so maybe walk through multiple times.
Then simplify after building graph. Refine by remove assembly parts not supported by PE.
RAM is a problem, optimal kmer is not known.

Question 5

How to improve de novo assemblies?

Answer 5

PE and MP, better coverage, hybrid methods.

Question 6

N50

Answer 6

Smallest contig in largest half of the assembly (calculate from assembly sum)

Question 7

kmer size’s influence on de Bruijn graph

Answer 7

Large k gives limited overlap, small gives a complex graph.