2012 NGS Field Guide – Table 4 – Advantages & Disadvantages of each instrument

Table 4. Primary advantages and disadvantages of each next-generation sequencing instrument.

Instrument Primary Advantages Primary Disadvantages
3730xl (capillary) Low cost for very small studies Very high cost for large amounts of data
454 GS Jr. Titanium Long read length; low capital cost; low cost per experiment High cost per Mb
454 FLX Titanium Long read length High cost per Mb
454 FLX+ Double the maximum read length of Titanium High capital cost; High cost per Mb; reagent issues; upgrade issues
PacBio Single molecule real-time sequencing; Longest available read length; Strobed reads; Short instrument run time; Low cost per sample; Many methods being developed High error rates; Low total number of reads per run; High cost per Mb; High capital cost; Many methods still in development; Weak company performance
Ion Torrent – PGM Low cost instrument upgraded through disposable chips (the chip is the machine); very simple machine with few moving parts; clear trajectory to improved performance Higher error rate than Illumina; more hands-on time relative to MiSeq
Ion Torrent – 314 chip Low cost per sample for small studies; Short time needed on instrument; Suitable for microbial sequencing and targeted sequencing; Easily upgraded with new chips High cost per Mb; Sample prep takes longer than time on the instrument
Ion Torrent – 316 chip Same as above, upgraded due to higher density chip Sample prep time
Ion Torrent – 318 chip Same as above, upgraded due to higher density chip; lower cost per read and Mb allows more applications Sample prep time
Ion Torrent – Proton Moderately low-cost instrument for high throughput applications Not yet publicly available; Error-rate likely higher than Illumina
Ion Torrent – Proton-I chip Same as PGM chips, but with many more sensors (wells) Not yet publicly available; Higher cost/Mb than HiSeq
Ion Torrent – Proton-II chip Same as Proton-I chip, but with many more sensors (wells); likely lower cost / Mb than HiSeq Not yet publicly available; Higher error-rate than Illumina
SOLiD – 5500xl Each lane of Flow-Chip can be run independently; High accuracy; Output in bases (not color-space); Ability to rescue failed sequencing cycles; 96 validated barcodes per lane; Throughput of 20-30Gb/day Not likely to be sold very long after the Ion Torrent Proton comes to market; Relatively short reads; more gaps in assemblies than Illumina data; less even data distribution than Illumina; High capital cost
Illumina MiSeq Low cost instrument and runs; Low cost/Mb for a small platform; Fastest Illumina run times and longest Illumina read lengths Relatively few reads and Higher cost/Mb compared to other Illumina platforms
Illumina HiScanSQ Versatile instrument for full catalog of Illumina arrays and sequencing; Scalable in future Higher cost/Mb than HiSeq for large amounts of data
Illumina GAIIx Lower Capital Cost than HiSeqs Slightly higher cost per Mb than HiSeq; Not as scalable in the future
Illumina HiSeq 1000 Lower instrument cost than HiSeq 2000; same number of reads/lane and cost/lane as HiSeq 2000; field upgradable to HiSeq 2000; Future scalability Not as flexible as HiSeq 2000 due to having only 1 flow cell
Illumina HiSeq 2000 Same as HiSeq 1000, but runs 2 flow cells simultaneously; Most reads, Gb per day and Gb per run, lowest cost per Mb of all platforms* High capital cost; High computation needs
Illumina HiSeq 2500 Same as HiSeq 2000, but can also run two 2 lane miniFlowCells to achieve much faster run times and longer read lengths miniFlowCell will likely have a higher cost per read than standard HiSeq Flow Cell; can’t run miniFlowCell and standard Flow Cell at the same time
Oxford Nanopore minION * No Instrument; IT IS A USB DEVICE; can load “raw” samples Not yet available; No data publicly available; High cost per Mb relative to other Nanopore sequencers
Oxford Nanopore GridION 2000 * Extremely long reads are feasible; Low-cost instrument (node); Nodes can placed in standard computer racks; Error-rate doesn’t increase along the length of the read; hairpin on one end allows reading of Not yet available; No data publicly available; 4% error-rates; errors are likely to be biased (thus multiple reads will lead to higher confidence in the wrong answer)
Oxford Nanopore GridION 8000 * Same as GridION 2000, but more reads per unit time; Lower cost per Gb Same as GridION 2000

Mb = Megabase; Gb = Gigabase; * Information based on company sources alone (independent data not yet available)