Contents
What you will do
This tutorial takes about 20 minutes and needs only a browser. You will open a seven-channel FISH image, give each channel its own lookup table, and use that view to check spot labels. The same steps work on your own multi-channel TIFFs.
Dataset · datatorch/ifish-chromosome-spotting-idr0123 · 50 files
iFISH chromosome spotting (IDR idr0123)
Seven-channel iFISH microscopy of human hTERT RPE-1 nuclei from IDR study idr0123 (Mota et al.), complete Dataset 5: 50 fields, each a maximum-intensity projection of the original 61-slice stacks with one page per channel (DAPI plus six DNA FISH probes AF488, AT542, AF594, AT647N, AF700, AF790 marking 23 loci along chromosome 1), 1024x1024, 16-bit, 0.13 um/px. Includes 18,754 published spot coordinates as 9x9 px bounding boxes in six fluorophore classes. Every channel can be shown, hidden, and window-leveled independently in the browser. License: CC BY 4.0. Source and original Z-stacks: https://idr.openmicroscopy.org/study/idr0123/ Please cite: Mota A, Schweitzer M, Wernersson E, Crosetto N, Bienko M. Simultaneous visualization of DNA loci in single cells by combinatorial multi-color iFISH. Scientific Data 9, 47 (2022).
Open →
The dataset
The hosted copy is Dataset 5 of IDR study idr0123 (Mota et al.): combinatorial multi-colour iFISH on human hTERT RPE-1 nuclei, with 23 loci along chromosome 1 marked by six fluorophores plus a DAPI counterstain.
| Fact | Value |
|---|---|
| Fields | 50, one TIFF per field, one page per channel |
| Channels | 7 |
| Image | 1024 x 1024 px, 16-bit, 0.13 µm per pixel |
| Projection | maximum-intensity projection of 61 Z slices |
| Labels | 18,754 spot boxes (9 x 9 px) in six fluorophore classes |
| License | CC BY 4.0 |
| Channel | Name in IDR | What it marks on chromosome 1 |
|---|---|---|
| DAPI | DAPI | counterstain (nucleus) |
| AF488 | a488 | 142.6 to 161.9 Mb |
| AT542 | tmr | 94.8 to 120.5 Mb |
| AF594 | a594 | 18.2 to 46.9 Mb |
| AT647N | Cy5 | 56.3 to 84.9 Mb |
| AF700 | a700 | 209.7 to 238.9 Mb |
| AF790 | ir800 | 171.7 to 200.3 Mb |
The original data is 3D. The hosted fields are projections, so a spot's Z position is not in the boxes. Per-dot fields such as nucleus ID, allele, intensity and SNR live in the source CSV on the Image Data Resource.
Step 1: look at one channel at a time
- 1
Open a field
Open the dataset above, click View all next to Files, open any field (for example field 001) and click Open in annotator. All seven channels load as separate layers in the Lookup Table tab.
- 2
Hide everything but DAPI
Click the Solo icon (the stacked layers) on the DAPI row. Every other channel turns off. You should see the nuclei and nothing else. This is your map of where spots can be.
- 3
Turn on one probe
Click Solo on the AF488 row to add it, then Solo on the DAPI row to take DAPI out, so AF488 is on alone. At default contrast many spots are faint, because the signal sits in a small part of the 16-bit range.
Step 2: give each channel its own lookup table
- 1
Window the faint channel
Click the AF488 row to select it. On the histogram above the rows, drag the LP (Low pass) handle left until the spots stand out, and the HP (High pass) handle right until the background between nuclei is close to black. Do not push it until the background turns to noise.
- 2
Repeat per channel
Do the same for each probe. Expect each channel to settle on a different window. That is the reason for per-channel settings: one setting for all seven would hide the dim channels or saturate the bright ones.
- 3
Bring DAPI back dim
Click Solo on the DAPI row last to bring it back, and drag its opacity slider down so nuclei give context without drowning the spots.
A lookup table changes the display, never the pixels. You can set an aggressive window to find spots and nothing in the file changes.
Step 3: check spot labels against the signal
Each spot box carries the label of its channel (AF488, AT542 and so on). Box colours come from the labels, not from the channel colours. With one probe channel on at a time:
- A box with no spot under it may be a detection that only shows in a Z slice the projection washed out. Note it and check the source stack.
- A spot with no box is either a missed detection or signal from a different channel bleeding through. Toggle the neighbouring channels to tell which.
- Boxes clustered tightly are expected: each fluorophore marks a span of chromosome 1 that holds 3 or 4 of the 23 loci.
Field 046 has no spots in the source table. It is a useful check that an empty field really is empty.
Step 4: do the same on your own images
Upload your own multi-channel TIFF to a project. Each page of the file becomes a channel you can show, hide and window on its own. If your files are OME-TIFF, the channel names stored in the metadata are shown, so you do not need to rename anything. Uncompressed 16-bit data is read at full depth and windowed down for display.
Source and license
IDR study idr0123, Dataset 5 (iFISH chr1 spotting, replicate 1). License: CC BY 4.0. Source and original Z-stacks: https://idr.openmicroscopy.org/study/idr0123/. Please cite: Mota A, Schweitzer M, Wernersson E, Crosetto N, Bienko M. Simultaneous visualization of DNA loci in single cells by combinatorial multi-color iFISH. Scientific Data 9, 47 (2022).
Open your own multi-channel images
Upload a multi-channel TIFF and set a lookup table per channel in the browser, with your team.
Frequently asked questions
Why does a multi-channel TIFF look black or washed out in an ordinary image viewer?
Most viewers stretch every channel the same way, or show only the first page. A 16-bit fluorescence channel often uses a small part of its range, so it needs its own window. A dim far-red channel and a bright DAPI channel almost never share good settings.
What is a per-channel lookup table?
A mapping from each channel's raw intensity to display brightness and colour. Setting it per channel lets you stretch a faint probe without blowing out the counterstain.
Does changing the lookup table change my data?
No. It changes only how the channel is displayed. The pixel values in the file, and any measurements taken from them, stay the same.
Can DataTorch read channel names from OME-TIFF?
Yes, channel names stored in OME-TIFF metadata are shown. The iFISH files in this tutorial are OME-TIFFs with one page per channel, so the channel panel shows the names in the table below.
Is the iFISH dataset free to use?
Yes. IDR study idr0123 is licensed CC BY 4.0. Cite Mota et al., Scientific Data 9, 47 (2022).
Open the hosted copy on DataTorch.
Browse the files in your browser, then download the dataset with a free account.
