Skip to content

seqplan: isolates per sequencing run

seqplan answers a question you have before sequencing: at the coverage I need, how many isolates fit on one run of this kit, how many runs do my samples need, and what does each isolate cost?

Sequencing is the expensive half of a genome project, and the price per isolate depends mostly on how many samples share a run. The arithmetic is simple, but it usually lives in a spreadsheet with kit outputs copied from somewhere. seqplan keeps each kit’s output next to the vendor page it came from, says whether a number is a specification or a reported typical yield, and adds a margin for uneven barcodes, which is the step people most often forget.

seqplan --kit miseq-v3 --genome-size 5.5M --coverage 50
seqplan -k nextseq-p1 -g klebsiella-pneumoniae -c 30 -n 96 --run-price 1500 --currency USD
seqplan -k miseq-v2 --reads 2x250 -s 2.8M -c 100
seqplan -k minion-r10.4.1 -g escherichia-coli -c 40 -n 24
seqplan --list-kits
seqplan --list-organisms
seqplan -k nextseq-p2 -s 5M --format json
Option Effect
-k, --kit ID Kit or flow cell (below). Required.
--reads 2x150 Read configuration. Default: the kit’s first listed one (2×300 for MiSeq v3, 2×150 for most others).
-g, --organism ID Organism preset with NCBI’s expected genome size.
-s, --genome-size SIZE Genome size in bp, with an optional k/M/G suffix (5.5M). Instead of --organism.
-c, --coverage X Target mean coverage. Default 50.
--usable F Fraction of the run’s output that ends up in sample reads. Default 0.8 for Illumina, 1.0 for nanopore (below).
--imbalance M Barcode-imbalance margin. Default 1.3 (below).
--barcodes B Barcodes your library kit offers (e.g. 24 or 96 for ONT Rapid Barcoding, 96 or 384 Illumina index pairs). Caps isolates per run, and the output says when it did.
-n, --samples N Isolates to sequence: adds runs needed and cost per isolate for them.
--run-price P, --currency C Price of one run in your currency. seqplan has no built-in prices.
-f, --format text|json Output format.
  • Per isolate: genome size × coverage, times the imbalance margin.
  • Per run: the kit’s output × the usable fraction.
  • Isolates per run: usable run output ÷ bases per isolate, rounded down. Both ends of the kit’s output range are reported.
  • Runs for N isolates: N ÷ isolates per run, rounded up. When one isolate needs more than a run (a large genome on a small kit), the runs are counted from the pooled bases instead.
  • Cost per isolate: run price ÷ isolates per run on a full run, or runs × price ÷ N for your N samples.

The two defaults are assumptions, not measurements. Change them to match your lab:

  • --usable 0.8 (Illumina). Illumina’s outputs are for a PhiX control library. Real runs lose some output to the PhiX spike-in, reads without a recognised barcode, and adapter and quality trimming. Nanopore yields below come from real runs already, so they default to 1.0.
  • --imbalance 1.3. Pooled libraries never split a run evenly, so each isolate is planned at 1.3× its share. The isolate that gets 1/1.3 of the average still reaches the target. Rapid nanopore barcoding can be much more uneven: in the study the MinION figures come from, one K. pneumoniae library got only 10.8 Mb and was rerun on its own flow cell.

Not modelled: library prep cost and the reads needed for anything other than coverage. Barcode set size is modelled only when you give --barcodes: a run that fits more isolates than you have barcodes is limited by the barcodes.

K. pneumoniae (5.5 Mb) at 50×, 96 isolates, for a run price of 1,500 USD (an example figure, not a quote):

Terminal window
$ seqplan -k miseq-v3 -s 5.5M -c 50 -n 96 --run-price 1500 --currency USD
Genome 5.5 Mb (given)
Kit MiSeq Reagent Kit v3, 2x300 (miseq-v3)
Coverage 50×
Per isolate 275 Mb at 50×; with the ×1.3 imbalance margin 357.5 Mb = 595.8 k 2x300 read pairs
Run output 13.2 to 15 Gb (vendor specification)
Usable 10.56 to 12 Gb (×0.8 for samples)
Isolates/run 29 to 33
Cost/isolate 45.45 to 51.72 USD on a full run
96 isolates 3 to 4 run(s), 46.88 to 62.50 USD per isolate
Source: https://www.illumina.com/systems/sequencing-platforms/miseq/specifications.html
Note: 22-25 M reads passing filter

At 25× the same run fits 59 to 67 isolates. On a NextSeq 2000 P1 (2×150, 30 Gb) it fits 67 at 50× and 134 at 25×. Halving the coverage roughly halves the sequencing cost per isolate. Whether 25× is enough for your analysis is a separate question, and one that should be measured.

Output per run (one flow cell), checked 2026-11-13. seqplan --list-kits prints the same table from the data compiled into the binary (seqplan/data/kits.tsv).

Id Kit Read configurations (output) Basis Source
miseq-v3 MiSeq Reagent Kit v3 2×300 (13.2–15 Gb), 2×75 (3.3–3.8 Gb) spec Illumina
miseq-v2 MiSeq Reagent Kit v2 2×150 (4.5–5.1 Gb), 2×250 (7.5–8.5 Gb), 2×25 (0.75–0.85 Gb) spec Illumina
miseq-v2-micro MiSeq Reagent Micro Kit v2 2×150 (1.2 Gb) spec Illumina
miseq-v2-nano MiSeq Reagent Nano Kit v2 2×150 (0.3 Gb), 2×250 (0.5 Gb) spec Illumina
iseq-100 iSeq 100 i1 Reagent v2 2×150 (1.2 Gb), 2×75 (0.6 Gb) spec Illumina
nextseq-550-mid NextSeq 500/550 Mid Output Kit v2.5 2×150 (32.5–39 Gb), 2×75 (16.25–19.5 Gb) spec Illumina
nextseq-550-high NextSeq 500/550 High Output Kit v2.5 2×150 (100–120 Gb), 2×75 (50–60 Gb) spec Illumina
nextseq-p1 NextSeq 1000/2000 P1 (XLEAP-SBS) 2×150 (30 Gb), 2×300 (60 Gb), 2×50 (10 Gb) spec Illumina
nextseq-p2 NextSeq 1000/2000 P2 (XLEAP-SBS) 2×150 (120 Gb), 2×300 (240 Gb), 2×100, 2×50 spec Illumina
nextseq-p3 NextSeq 2000 P3 (XLEAP-SBS) 2×150 (360 Gb), 2×100, 2×50 spec Illumina
nextseq-p4 NextSeq 2000 P4 (XLEAP-SBS) 2×150 (540 Gb), 2×100, 2×50 spec Illumina
novaseq-x-1.5b NovaSeq X 1.5B 2×150 (500–716 Gb), 2×300, 2×100, 2×50 spec Illumina
novaseq-x-10b NovaSeq X 10B 2×150 (3–4 Tb), 2×100, 2×50 spec Illumina
novaseq-x-25b NovaSeq X 25B 2×150 (8–10.5 Tb), 2×100, 2×50 spec Illumina
minion-r10.4.1 MinION flow cell R10.4.1, rapid barcoding long reads (5.2–7.0 Gb) typical Microbial Genomics 2024
minion-r10.4.1-core MinION flow cell R10.4.1, core facility long reads (10–30 Gb) typical UGA GGBC
flongle-r10.4.1 Flongle flow cell R10.4.1 long reads (1.8–2.6 Gb) typical University of Iowa
  • Illumina figures are the vendor’s specifications for a PhiX control library at supported cluster densities. For NovaSeq X, Illumina doesn’t guarantee the upper end. The NextSeq 500/550 kits are listed on Illumina’s NextSeq 550 page. The NextSeq 1000/2000 figures are for XLEAP-SBS chemistry. P3 and P4 run on the NextSeq 2000 only.
  • Nanopore yields depend on the library, the run time and the sample far more than Illumina’s do, so seqplan uses reported typical ranges, not ONT’s theoretical maxima.
    • The MinION rapid-barcoding range comes from 72-hour runs of bacterial isolates in a published study.
    • The core-facility MinION and Flongle ranges are what sequencing cores state they expect.
    • Oxford Nanopore’s own figures are theoretical maxima: up to 50 Gb for a MinION flow cell over 72 h (price list), and up to 2.8 Gb for a Flongle (specification).
    • If your own runs yield less than the listed range, scale it down with --usable (your yield ÷ the listed yield).

Genome sizes from NCBI’s expected-genome-size service (the size NCBI checks submitted assemblies against), fetched 2026-11-13. Any other organism: give --genome-size.

Id Expected size NCBI range
klebsiella-pneumoniae 5.6 Mb 4.5–6.7 Mb
escherichia-coli 5.15 Mb 4.0–6.3 Mb
salmonella-enterica 5.05 Mb 4.1–6.0 Mb
staphylococcus-aureus 2.8 Mb 2.2–3.4 Mb
pseudomonas-aeruginosa 6.475 Mb 5.0–7.95 Mb
acinetobacter-baumannii 3.95 Mb 3.2–4.7 Mb
enterococcus-faecium 2.95 Mb 2.0–3.9 Mb
enterococcus-faecalis 2.95 Mb 2.0–3.9 Mb
streptococcus-pneumoniae 2.095 Mb 1.57–2.62 Mb
mycobacterium-tuberculosis 4.075 Mb 3.15–5.0 Mb
vibrio-cholerae 4.225 Mb 3.6–4.85 Mb
neisseria-gonorrhoeae 2.225 Mb 1.95–2.5 Mb
helicobacter-pylori 1.62 Mb 1.29–1.94 Mb
campylobacter-jejuni 1.985 Mb 1.37–2.6 Mb
listeria-monocytogenes 3.05 Mb 2.5–3.6 Mb
bacillus-subtilis 4.12 Mb 3.3–4.95 Mb

The seqplan tool takes the same options and returns the plan as JSON, so an AI agent can answer “how many isolates per run” from the cited figures instead of from memory. It never fills in a price the user didn’t give.