Illumina Instrument Type from fastq?

Question

[andrew.j.skelton73]

Hi,

Any idea if there's a way to get the Instrument type (i.e. NextSeq, HiSeq, etc), from the instrument name field in the Fastq read header? - I've tried looking for a map of sorts in Illumina's documentation, but to no avail.

Thanks,

[mmulinge]

Biostars.org is my one stop solution for bioinformatics info and analysis software Keep it up.

[Olov]

Hi!

Anybody knows what the instrument is for this header?

@NB552493

[GenoMax]

That is either a NextSeq 500 or 550.

[Olov]

Awesome thanks

Answer 1

[GenoMax]

There was a post @SeqAnswers where we had a recap of the types of sequencer specific fastq headers (my google-fu has not turned up that post yet). Following is a very rough approximation of start of fastq headers that are sequencer specific.

Edit: Just want to add that it is possible to change the default read header in one of the set up files (so the following may not always be true).

@HWI-Mxxxx or @Mxxxx - MiSeq
@HWUSI - GAIIx
@HWI-Dxxxx - HiSeq 2000/2500
@Kxxxx - HiSeq 3000(?)/4000
@Nxxxx - NextSeq 500/550
@Axxxxx - NovaSeq
@Vxxxxx = NextSeq 2000
Edit (08/2022)
@AAxxxxx - NextSeq 2000 P1/P2/P3
@Hxxxxxx - NovaSeq S1/S2/S4

[igor]

Actually, the MiSeq read headers should be @Mxxxx

[harold.smith.tarheel]

I don't know if those are still the default headers. Our new MiSeq outputs '@Mxxx' headers so, unless the FAS reconfigured it upon installation, it's the current default. Similarly, our HiSeq headers are not prefaced with '@HISEQ'.

The FCIDs are generated automatically during the flow cell scan, which is why I recommended using those to unambiguously distinguish the instruments. But I have not yet been able to find the complete code for conversion. I'll continue digging.

[GenoMax]

That is why I added a disclaimer.

It is possible to reconfigure the default headers (we have several HiSeq/MiSeq and the headers are slightly different). I have not looked through the config files to see where exactly this is set.

I remember that the FAS's have access to a special illumina tool/database where they can lookup kit/flowcell ID's but I don't think that tool is available to us.

[GenoMax]

Do you mean FCID = Name of the data folder?

That generally has the format

For HiSeq

YYMMDD_InstrumentSerialNumber_RunNumber*_[A/B]FlowCellBarcode

For MiSeq

YYMMDD_InstrumentSerialNumber_RunNumber*_000000000-FlowCellBarcode

YYMMDD - Date the run started
*RunNumber - Cumulative run number (incremented by 1 for each run) on that sequencer.
A/B - signifies the flowcell position (in case of instruments that can run 2 FC at a time).

These names are configurable in a .cfg settings file and will give one an idea of the kind of sequencer it is (provided the default naming scheme is left intact).

[harold.smith.tarheel]

FCID = flow cell identifier (same as your FlowCellBarcode). In the read identifier, it's the third field.

[andrew.j.skelton73]

Yes! That's what I was looking for, and I vaguely remember seeing the same post on seq answers too! This post maybe?

[GenoMax]

That is one of them. I seem to recollect that there was another one specifically like the question you asked.

[Devon Ryan]

HiSeq 3000 starts with a J, not a K. Also, our HiSeq 2500 defaults to SNXXXXX.

Edit: Apparently HiSeq 3000s can start with either (consistency!). I bet our HiSeq 2500 starts with an S because it started life as a 2000 and got an upgrade.

Answer 2

[Devon Ryan]

I realize that this is a year and a half late, but here is some code from 10X that incorporates both the flow cell ID and the machine ID to figure out the run type.

[andrew.j.skelton73]

Fantastic find - really useful.

[Devon Ryan]

I wonder if anyone has NovaSeq data so we could add that info here...

[GenoMax]

Not 100% sure: S2 cells "H[A-Z,0-9]{4}DMXX$"
Sequencer serial may be "A[0-9]{6}$" for NovaSeq 6000.

[Dan D]

NovaSeq S4 flowcell barcodes are ^H[A-Z,0-9]{4}DSXX$

[Devon Ryan]

Well that was quick (I notice now that 10X has a line for the NovaSeq in the source code I linked to).

[GenoMax]

Guess that number is for S2 FC. So someone will have to give us data for S1/S4.

[Dan D]

Every NovaSeq serial I've seen has been of the form ^A\d{5}$ (five, not six digits). That's just at one site though.

[andrew.j.skelton73]

Still doing calibration runs on the nova at my institute, but I'll update when I get some data through

[ajay nair]

Here are the flowcell barcodes for NovaSeq 6000. For every type: [A-Za-z0-9]{5}D[SMR]{1}X[XY2357]

[i.sudbery]

Anyone know what a Instrument ID starting "@ST-E" means?

[GenoMax]

Is this recent data? Could it be from a iSeq 100?

[i.sudbery]

I doubt it, there is 30M reads in each sample.

[GenoMax]

True. Do you have the FC barcode?

[mvizue01]

It's a HiSeqX, If you check some of Illumina's public data from one of those instruments you will find that the instrument ID will either start with a E or an ST-E.

Answer 3

[harold.smith.tarheel]

Illumina provided the following information regarding FCID mapping to instruments. The indicated characters are the last four of the nine-character FCID (excepting MiSeq). Please note that this is not comprehensive and there may be new additions in the future.

AAXX = Genome Analyzer
BCXX = HiSeq v1.5
ACXX = HiSeq High-Output v3
ANXX = HiSeq High-Output v4
ADXX = HiSeq RR v1
AMXX, BCXX =HiSeq RR v2
ALXX = HiSeqX
BGXX, AGXX = High-Output NextSeq
AFXX = Mid-Output NextSeq
5 letter/number = MiSeq

[GenoMax]

Thank you for getting this from Illumina.

While it does not conclusively associate a run with the kind of sequencer used (I think some of the HiSeq flowcells can be used on multiple models) it does provide useful information. Also the chances of flowcell ID being altered (except with SRA :-)) are small so this association should be more reliable.

[steven.davis]

What kind of instrument uses flowcells ending with BBXX? Here is the read sequence id line:

@ERR1417747.1 K00150:65:H77FVBBXX:7:1101:1570:1297 length=150

[GenoMax]

The first part @ERR1417747 is SRA accession ID that get's tacked on (unless you use -F to regenerate sequence headers in the original Illumina format). Considering the sequencer ID K00150 that follows this should be a HiSeq 3000/4000 run.

[liangzebin5566]

I got some fastq files, the id is like:

@FCH7HCYADXY:1:1101:10595:1825

what kind of instrument can it be? How can I get the latest RCID description?

Answer 4

[harold.smith.tarheel]

I believe you can extract the instrument type from the FCID in the read identifier (e.g. our HiSeq FCIDs all end in 'ACXX', MiSeqs all start with 'MS'), but you may need to contact Illumina to determine the complete code.

[andrew.j.skelton73]

This was along the lines of what I was thinking, as there are similar tags in all the nextSeq data I see coming through, however I was hoping somebody had a handy map of these in some dark corner of the internet I've not come across yet!

Answer 5

[Nathan]

This page from Illumina also has some helpful information regarding flowcell serial numbers. The following R snippet may also be of use (it's just concatenated from the information mentioned here and above):

  mutate(
    sn_infer = case_when(
      str_detect(sn, "^HWI-M|^M") ~ "MiSeq",
      str_detect(sn, "^HWUSI") ~ "GAIIx",
      str_detect(sn, "^HWI-D") ~ "HiSeq 2x00",
      str_detect(sn, "^K") ~ "HiSeq 3/4000",
      str_detect(sn, "^N") ~ "NextSeq 5x0",
      str_detect(sn, "^A|^H") ~ "NovaSeq",
      str_detect(sn, "^V|^AA") ~ "NextSeq 2000",
    ),
    fc_infer = case_when(
      str_detect(fc, "^(BRB|BP[ACGL]|BNT)") ~ "iSeq_100",
      str_detect(fc, "000H") ~ "MiniSeq",
      str_detect(fc, "^[BCJKDG]") ~ "MiSeq",
      str_detect(fc, "(A[FG]|BG)..$") ~ "NextSeq_500/550 ",
      str_detect(fc, "M5$") ~ "NextSeq_1000/2000",
      str_detect(fc, "HV$") ~ "NextSeq_2000",
      str_detect(fc, "([AB]C|AN)..$") ~ "HiSeq_2500",
      str_detect(fc, "BB..$") ~ "HiSeq_3000/4000",
      str_detect(fc, "(AL|CC)..$") ~ "HiSeq_X",
      str_detect(fc, "D[RS]..$|DM.$") ~ "NovaSeq_6000"
    )

[GenoMax]

The link included in this answer no longer appears to be working.

Answer 6

[nickp60]

I have run into this question several times over the years, and wrote this package to provide a command line tool for this problem. PRs and other feedback welcome! https://github.com/nickp60/fcid

$ fcid 22C37GLT3
NovaSeq X,NovaSeq X Plus
$ fcid MN10009 --by-machine
MiniSeq