An Introduction to Genome Assembly

Contributors

Authors: Simon Gladman

Questions

• How do we perform a very basic genome assembly from short read data?

Objectives

• assemble some paired end reads using Velvet

• examine the output of the assembly.

.enlarge120[

De novo Genome Assembly

]

With thanks to T Seemann, D Bulach, I Cooke and Simon Gladman

.enlarge120[

De novo assembly

]

.pull-left[

The process of reconstructing the original DNA sequence from the fragment reads alone.

• Instinctively like a jigsaw puzzle

  • Find reads which “fit together” (overlap)
  • Could be missing pieces (sequencing bias)
  • Some pieces will be dirty (sequencing errors)

]

.pull-right[ ]

Another View

Assembly: An Example

A small “genome”

Shakespearomics

Shakespearomics

Shakespearomics

So far, so good!

The Awful Truth

“Genome assembly is impossible.” - A/Prof. Mihai Pop

.enlarge120[

Why is it so hard?

]

.pull-left[

• Millions of pieces
  • Much, much shorter than the genome
  • Lots of them look similar
• Missing pieces
  • Some parts can’t be sequenced easily
• Dirty Pieces
  • Lots of errors in reads ]

.pull-right[ ]

Assembly recipe

• Find all overlaps between reads
  • Hmm, sounds like a lot of work..
• Build a graph
  • A picture of the read connections
• Simplify the graph
  • Sequencing errors will mess it up a lot
• Traverse the graph
  • Trace a sensible path to produce a consensus

A more realistic graph

.image-15[] What ruins the graph?

• Read errors
  • Introduces false edges and nodes
• Non haploid organisms
  • Heterozygosity causes lots of detours
• Repeats
  • If they are longer than the read length
  • Causes nodes to be shared, locality confusion.

Repeats

.enlarge120[

What is a repeat?

]

.pull-left[

A segment of DNA which occurs more than once in the genome sequence

• Very common
  • Transposons (self replicating genes)
  • Satellites (repetitive adjacent patterns)
  • Gene duplications (paralogs)

]

.pull-right[

]

Effect on Assembly

.enlarge120[

The law of repeats .image-15[]

]

It is impossible to resolve repeats of length S unless you have reads longer than S

It is impossible to resolve repeats of length S unless you have reads longer than S

Scaffolding

.enlarge120[

Beyond contigs

]

.pull-left[

Contig sizes are limited by:

• the length of the repeats in your genome
  • Can’t change this
• the length (or “span”) of the reads
  • Use long read technology
  • Use tricks with other technology

]

.enlarge120[

Types of reads

]

.pull-left[.enlarge120[Example fragment]]

.remark-code[.enlarge120[atcgtatgatcttgagattctctcttcccttatagctgctata]]

.pull-left[.enlarge120[“Single-end” read]]

.remark-code[.enlarge120[atcgtatgatcttgagattctctcttcccttatagctgctata]]

sequence one end of the fragment

.pull-left[.enlarge120[“Paired-end” read]]

.remark-code[.enlarge120[atcgtatgatcttgagattctctcttcccttatagctgctata]]

sequence both ends of the same fragment

We can exploit this information!

.enlarge120[# Scaffolding]

• Paired end reads
  • Known sequences at each end of fragment
  • Roughly known fragment length

• Most ends will occur in same contig

• Some will occur in different contigs

  • evidence that these contigs are linked

.enlarge120[# Contigs to Scaffolds]

.enlarge120[# Assessing assemblies]

• We desire
  • Total length similar to genome size
  • Fewer, larger contigs
  • Correct contigs
• Metrics
  • No generally useful measure. (No real prior information)
  • Longest contigs, total base pairs in contigs, N50, …

.enlarge120[# The “N50”]

.enlarge120[The length of that contig from which 50% of the bases are in it and shorter contigs]

• Imagine we have 7 contigs with lengths:
  • 1, 1, 3, 5, 8, 12, 20
• Total
  • 1+1+3+5+8+12+20 = 50
• N50 is the “halfway sum” = 25
  • 1+1+3+5+8+12 = 30 (>25) so N50 is 12

.enlarge120[# 2 levels of assembly]

• Draft assembly
  • Will contain a number of non-linked scaffolds with gaps of unknown sequence
  • Fairly easy to get to
• Closed (finished) assembly
  • One sequence for each chromosome
  • Takes a lot more work
  • Small genomes are becoming easier with long read tech
  • Large genomes are the province of big consortia (e.g. Human Genome Consortium)

.enlarge120[# How do I do it?] — .enlarge120[

Example

• Culture your bacterium

• Extract your genomic DNA

• Send it to your sequencing centre for Illumina sequencing
  • 250bp paired end
• Get back 2 files
  • .remark-code[MRSA_R1.fastq.gz]
  • .remark-code[MRSA_R2.fastq.gz]
• Now what? ]

.enlarge120[# Assembly tools

• Genome
  • Velvet, Velvet Optimizer, Spades, Abyss, MIRA, Newbler, SGA, AllPaths, Ray, SOAPdenovo, …
• Meta-genome
  • Meta Velvet, SGA, custom scripts + above
• Transcriptome
  • Trinity, Oases, Trans-abyss

And many, many others…

]

.enlarge120[

Assembly Exercise #1

• We will do a simple assembly using Velvet in Galaxy
• We can do a number of different assemblies and compare some assembly metrics.

]

Key Points

• We assembled some Illumina fastq reads into contigs using a short read assembler called Velvet
• We showed what effect one of the key assembly parameters, the k-mer size, has on the assembly
• It looks as though there are some exploitable patterns in the metric data vs the k-mer size.

Thank you!