Work Log 01/24 - Speaker Deck

Slide 1

Slide 1 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University 2014.01 Slides by Liang Bo Wang

Slide 2

Slide 2 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University 2014.01 Slides by Liang Bo Wang

Slide 3

Slide 3 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University 2014.01 Slides by Liang Bo Wang 10M + 1M each 99K 250K 740K list price New Lineup of Illumina Sequencers

Slide 4

Slide 4 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University HiSeq X Ten •  $1,000 genome reached –  include typical instrument depreciation, DNA extraction, library preparation, and estimated labor •  bundled as at least 10 HiSeq X machines –  new optics and chemistry makes them run 10x faster than HiSeq 2500, 18TB in 3 days w/ 6B clusters –  new flowcells (use nanowells) 2014.01 Slides by Liang Bo Wang

Slide 5

Slide 5 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University Cost Behind the $1,000 Holy Grail •  $ 10M capital budget for machines only •  Run all machines 24/7/365 for 4 years “to get the instrument amortization costs down to $135 per genome, they stretched out the lifecycle to four years” •  You’ve got 72,000 human whole genome samples •  Requires $ 67M operating cost during 4 years –  library prep = reagents AND labor = $65 per sample •  Who could analyze the data? 2014.01 Slides by Liang Bo Wang

Slide 6

Slide 6 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University Who has bought HiSeq X Ten? •  one set + 4 by Broad Institute •  one set by the Garvan Institute of Medical Research from Australia •  one set by Macrogen, leading next-generation sequencing service organization based in Seoul, South Korea and its CLIA laboratory in Rockville, Maryland 2014.01 Slides by Liang Bo Wang

Slide 7

Slide 7 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University NextSeq 500 •  130M / 400M clusters per run –  120 Gb with 150bp pair end •  Somewhere between HiSeq and MiSeq •  New optics allow six cameras in a single unit one third of the cost of the current model, and now use an LED allowing Illumina –  … ? 2014.01 Slides by Liang Bo Wang

Slide 8

Slide 8 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University Current(Previous) Technology •  two-color laser, 4 bases with separate dyes. •  a filter wheel to discriminate the spectra •  4 pictures are captured by CCD per SBS cycle 2014.01 Slides by Liang Bo Wang

Slide 9

Slide 9 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University Tech used by HiSeq 500 (Assumed) •  only two dyes are used –  two based labeled with single dye –  third with both dyes –  fourth with no dyes •  only two pictures are taken per SBS cycle –  make computation easier –  low lib prep complexity –  reagent and instrument will be cheaper 2014.01 Slides by Liang Bo Wang

Slide 10

Slide 10 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University 2014.01 Slides by Liang Bo Wang MiSeq NextSeq 500 HiSeq 2500 Throughput (Gb) 0.5 - 15 20 - 39 / 30 - 120 10 - 180 / 50 - 1000 Run time 5 - 65 (h) 15 - 26 / 12 - 30 (h) 7 - 40h / <1 - 6d DNA-Seq Whole Genome (human 30x) - 1 1 - 10 Exome, small panel 3 15 - 48 36 - 72 RNA-Seq Transcriptome - 3 - 10 8 - 96 sRNA 1 - 5 25 - 80 60 - 792 Methylation ChIP-Seq 1 8 - 24 20 - 264 Methylation (30x) - 1 1 - 10 adapted from Illumina datasheet

Slide 11

Slide 11 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University Reference •  http://core-genomics.blogspot.tw/2014/01/illuminas-christmas-presents.html •  http://biomickwatson.wordpress.com/2014/01/15/illumina-destroy-the- opposition-again-almost/ •  http://blog.allseq.com/1000-genome-72m •  http://nextgenseek.com/2014/01/what-is-the-price-of-nextseq-500-and-hiseq- x-ten/ •  http://nextgenseek.com/2014/01/how-does-nextseq-500-compare-with- miseq-and-hiseq/ 2014.01 Slides by Liang Bo Wang

Slide 12

Slide 12 text

Bioinformatics and Biostatistics Core, NTU Center of Genomic Medicine, National Taiwan University Future Work •  Test new RNA-Seq pipeline –  Tophat + Cufflinks cannot be scaled on Hadoop –  STAR, MapSplice, … •  Run DNA-Seq GATK 2.x pipeline –  GATK 1.x outdated, parameters change •  Analysis report form –  Can also be used for Phalanx service •  UI design (for prototype) 2014.01 Slides by Liang Bo Wang