Skip to content

Latest commit

 

History

History
406 lines (277 loc) · 11.7 KB

tigmint-slides.md

File metadata and controls

406 lines (277 loc) · 11.7 KB
Raw
author date history slideNumber
Shaun D Jackman
2017-07-24
true
true

Tigmint

Correct Misassemblies Using Linked Reads From Large Molecules

Shaun Jackman @sjackman

Benjamin P Vandervalk, Rene L Warren, Hamid Mohamadi, Justin Chu, Sarah Yeo, Lauren Coombe, Joerg Bohlmann, Steven JM Jones, Inanc Birol

| ISMB HiTSeq 2017-07-25 | https://sjackman.ca/tigmint-slides | Creative Commons Attribution License

Shaun Jackman

| BCCA Genome Sciences Centre | Vancouver, Canada | @sjackman · github.com/sjackman · sjackman.ca

https://github.com/bcgsc/tigmint

10x Genomics Chromium Linked Reads http://www.10xgenomics.com/assembly/

+ Large molecules of DNA (100 kbp and up) are isolated in partitions + Each partition has its own barcode + Reads from the same molecule share the same barcode + One nanogram of input DNA with 10x Genomics Chromium

Linked Reads

  • Call variants in repetitive regions
  • Identify structural variants
  • Phase variants

Tools for Linked Reads

10x Genomics

Long Ranger · Supernova

Structural Variants

GROC-SVs · Topsorter

Scaffolding

ARCS · Architect · Fragscaff

+ [Architect][] intended for synthetic long reads + [Fragscaff][] intended for contiguity-preserving transposition + [Topsorter][] requires a VCF file

| B-040 · Local assembly with Kollector · Justin Chu | B-048 · ABySS 2.0 · Ben Vandervalk

+ One sequencing library + Scaffold over gaps and repeats + Local assembly of gaps and repeats + Identify and correct misassemblies

Contigs and scaffolds come to an end due to...

  • repeats
  • sequencing gaps
  • heterozygous variation
  • misassemblies

Misassemblies limit contiguity

particularly for highly contiguous assemblies.

Most scaffolding algorithms address repeats and gaps, but not misassemblies.

Correct misassemblies

Correct misassemblies

Scaffold

Tigmint

  • Count molecules that support and refute each position of the assembly
  • Output positions of possible misassemblies

  • Regions with poor molecule depth are suspect
  • Clipped read alignments at molecule ends
    pinpoint the position of the breakpoint
  • Highlight possible misassemblies

Scatter plot of molecule start and end coordiantes

Graph of 10 kbp segments sharing barcodes

Menagerie of Misassemblies

  • Chimeric sequence
  • Missing sequence (deletion)
  • Chimeric insertion
  • Inversion
  • Collapsed repeat

ABySS 2.0 assembly of GIAB HG004

Tigmint, at least 4 clipped molecules

Tigmint, at least 3 clipped molecules

Tigmint, at least 2 clipped molecules

Sitka Spruce Mitochondrion

16 scaffolds, N50 is 0.46 Mbp

Sitka Spruce Mitochondrion

Before and After Tigmint

Before Tigmint After Tigmint + ARCS

| N50 improved nine fold from 0.46 Mbp to 4.2 Mbp | Poster B-066

+ Assembled organelles from WGSS \ with ABySS 2.0, ARCS, and Tigmint + 6 Mbp mitochondrial genome in 4 scaffolds > 50 kbp + Nine fold improvement in N50 + 70% of genome in one 4.2 Mbp scaffold
Tools Scaffolds N50
ABySS + ARCS 16 0.46 Mbp
+ Tigmint 4 4.2 Mbp

fin

Shaun Jackman

| BCCA Genome Sciences Centre | Vancouver, Canada | @sjackman · github.com/sjackman · sjackman.ca

Tigmint
https://github.com/bcgsc/tigmint

Slides
https://sjackman.ca/tigmint-slides

Markdown source code
https://github.com/sjackman/tigmint-slides

Links

ABySS · ARCS · Architect · Fragscaff · GROC-SVs · LINKS · Long Ranger · MAKER · Pilon · Prokka · Sealer · Supernova · Tigmint · Topsorter

Supplementary Slides

Linked reads

  • Large molecules of DNA (100 kbp and up) are isolated in partitions
  • Each partition has its own barcode
  • Reads from the same molecule share the same barcode
  • One nanogram of input DNA with 10x Genomics Chromium

Scaffold with ARCS

  • Map the reads to the assembly
  • Identify scaffold ends sharing common barcodes
  • Construct a graph
  • Each edge connects two scaffold ends
  • Merge unambiguous paths through this graph

ARCS: Assembly Roundup by Chromium Scaffolding
bioRxiv: https://doi.org/10.1101/100750

ARCS https://github.com/bcgsc/arcs

Visualization

  • Graph of 10 kbp segments sharing barcodes
  • Scatter plot of molecule start and end position
  • Physical molecule coverage
  • Histogram of clipped read alignments at molecule ends
  • Highlight breakpoints (or putative misassemblies)

Chimeric scaffold involving a 2 kbp repeat

Chimeric sequence

Human Assembly

  • Assemble paired-end/mate-pair with ABySS 2.0
  • Correct misassemblies with 10x Chromium
  • Assembly locus discordant with both the linked reads and the reference is likely a misassembly

Requiring 4 clipped molecules
Tigmint identified 39 breakpoints.
38 of 39 (97%) discordant with the reference

Requiring 3 clipped molecules,
Tigmint identified 184 breakpoints.
104 of 184 (57%) discordant with the reference

Requiring 2 clipped molecules,
Tigmint identified 1995 breakpoints.
250 of 1995 (12.5%) discordant with the reference

Human Assemblies

Assembly Tigmint breaks Discordant breaks Breaks before Breaks after
ABySS 39 38 (97%) 2717 2679
ABySS + Bionano + ARCS 78 14 (18%) 2757 2743
Supernova 163 56 (34%) 3883 3827

Genome in a Bottle HG004

Genome Skimming

Assemble the 6 Mbp Sitka spruce mitochondrion

  • Whole genome sequencing data contains both nuclear and organellar reads
  • Hundreds of mitochondria and plastids per cell
  • Reads of the organellar genomes are abundant
  • Organellar genomes assemble with a single lane
  • Single-copy nuclear sequences are too low depth to assemble well

White spruce depth vs percent GC

Method

Sitka Spruce Mitochondrion

  • Assembled organelles from WGSS
    with ABySS 2.0, ARCS, and Tigmint
  • 6 Mbp mitochondrial genome in 4 scaffolds
  • Nine fold improvement in N50
  • 70% of genome in one 4.2 Mbp scaffold
Tools Scaffolds N50
ABySS + ARCS 16 0.46 Mbp
+ Tigmint 4 4.2 Mbp

Depth of coverage per scaffold

Sitka spruce mitochondrion

Read Metrics Plastid Mitochondrion
Number of HiSeq lanes 1 GemCode lane 1 Chromium lane
Read length 2 x 125 bp 2 x 150 bp
Number of read 630 million 843 million
Number selected for assembly 4.3 million 119 million
Number mapped to assembly 15,232 of 4.3 M 3.78 M of 843 M
Proportion of organellar reads 1/283 0.35% 1/223 or 0.45%
Depth of coverage 17x 40x
Assembly Metrics Plastid Mitochondrion
Assembled genome size 124,029 bp 6.09 Mbp
Number of contigs 1 contig 1,216 contigs
Contig N50 124 kbp 13.7 kbp
Number of scaffolds 1 scaffold 239 scaffolds
Scaffold N50 124 kbp 461 kbp
Largest scaffold 124 kbp 1,223 kbp
GC content 38.8% 43.6%
Annotation Metrics Plastid Mitochondrion
Number of genes w/o ORFS 114 (108) 115 (67)
Protein-coding genes (mRNA) 74 (72) 84 (47)
rRNA genes 4 (4) 3 (2)
tRNA genes 36 (32) 25 (18)
ORFs ≥ 300 bp 4 1,154
Introns in coding genes 9 (8) 9 (6)
Introns in tRNA genes 6 (6) 0