How to Karyotype: A Chromosome-by-Chromosome Guide for Trainees
Published 2026-09-13
Nobody learns to karyotype by reading about it.
You learn it the way you learned faces. The first week, twenty-three pairs of grey smudges all look like the same grey smudge. Somewhere around week three a 9 walks past and you know it is a 9 before you have consciously checked anything, and you cannot really explain how. That is pattern recognition doing what pattern recognition does, and the only way to get there is repetition on real chromosomes.
So this guide is built to be followed with your hands, not just read. The video below is a complete run through a metaphase spread, chromosome by chromosome, in a free browser tool. Open the tool in a second tab and rebuild the same karyogram while you watch.
[[video-how-to-karyotype]]
Follow along here: CytoSnap is free, runs in the browser, and needs no install. Case 1 is open to everyone with no account.
Chapters: 0:33 a finished example · 1:04 chromosomes 1 to 5 · 2:22 chromosomes 6 to 12 · 4:00 the acrocentrics · 4:29 chromosomes 16 to 18 · 5:18 chromosomes 19 to 22 · 5:55 the X and the Y · 6:21 placing them three different ways · 8:39 flip and rotate
What you are actually doing
A karyotype is not a picture of chromosomes. It is a picture of chromosomes sorted into a standard order so that a human eye can find the one thing that is wrong.
The order is not arbitrary and it is not yours to change. Chromosomes are laid out largest to smallest, 1 through 22, then the sex chromosomes. Every chromosome sits with its short arm up. Short arm is p, from the French petit; long arm is q, which is simply the next letter. Homologues sit side by side with their centromeres on the same horizontal line, which is the whole point of the layout: when two partners are lined up at the centromere, a difference in arm length or banding pattern jumps out instead of hiding.
That finished, sorted image is a karyogram. The karyotype is the description of it, written in ISCN: 46,XX, or 46,XY,t(9;22)(q34;q11.2). People use the two words interchangeably and nobody will correct you at the bench, but the distinction is worth knowing because the exam draws it.
Before any of this, the cell had to be worth sorting. A metaphase you can karyotype has minimal overlap, chromosomes that are neither pulverised nor fuzzy, and a banding level you can actually resolve. If you are not confident about that part yet, read how to read chromosome banding first, because sorting a bad spread teaches you bad habits.
The three questions
Every chromosome you pick up gets the same three questions, in this order. Beginners skip straight to question three and drown. Do not.
1. How big is it? Not in microns. Relative to its neighbours on the screen. Size alone gets you into the right neighbourhood within about a second, and the neighbourhoods have names (below).
2. Where is the centromere? This is the pinch, the point where the two chromatids are joined and the outline narrows. Three answers:
| Centromere position | Name | What it looks like | Examples |
|---|---|---|---|
| Near the middle | Metacentric | p and q arms roughly equal | 1, 3, 16, 19, 20 |
| Off centre | Submetacentric | Clearly short p, clearly long q | 2, 4, 5, 6 to 12, 17, 18, X |
| Right at the top | Acrocentric | p arm is a stub with satellites | 13, 14, 15, 21, 22, Y |
The acrocentric p arms carry ribosomal DNA and stalk-and-satellite material. They are polymorphic, they vary in length between people, and you do not report them as abnormal when they look big or small. They are also why Robertsonian translocations happen almost exclusively between these five chromosomes.
3. What is the banding pattern? Now, and only now, you read the stripes. Dark G bands are AT-rich, late-replicating and gene-poor. Pale bands are GC-rich, early-replicating and gene-dense. The pattern along each arm is reproducible enough to be a fingerprint, which is the entire reason karyotyping works.
The seven groups
The A to G grouping predates banding. It is how chromosomes were sorted when all you had was size and centromere position, and it survives because it is still how your eye triages a field.
| Group | Chromosomes | Defining feature |
|---|---|---|
| A | 1, 2, 3 | Largest. 1 and 3 metacentric, 2 submetacentric |
| B | 4, 5 | Large submetacentric, very short p arms |
| C | 6 to 12, X | Medium submetacentric. The crowded, difficult group |
| D | 13, 14, 15 | Medium acrocentric |
| E | 16, 17, 18 | Shorter. 16 metacentric, 17 and 18 submetacentric |
| F | 19, 20 | Short metacentric, notably pale |
| G | 21, 22, Y | Shortest acrocentric |
Group C is where trainees lose time, because seven autosomes plus the X sit in a narrow size range with similar centromere positions. That is the group worth drilling hardest.
Chromosome by chromosome
What follows are the landmarks that survive at a working banding level of roughly 400 to 550 bands, which is what most clinical spreads give you. Some of the mnemonics are silly. Silly is the point: the ridiculous ones are the ones that come back to you at 4pm on a Friday.
Group A: 1, 2, 3
- 1 is the biggest thing in the field and metacentric. Its tell is the large dark block immediately below the centromere on q (the 1q12 heterochromatin). That block is polymorphic and varies between people, so a long one is not a finding.
- 2 is the largest submetacentric chromosome: a clearly shorter p, a clearly longer q, and bands spaced fairly evenly down both arms. Nothing dramatic, which is itself the clue.
- 3 is metacentric with arms close to equal length, and it has the most distinctive centromere in the karyotype: a pale band straddling the centromere with dark bands on both sides, like a light stripe sandwiched in the middle. Once you see the sandwich you never confuse a 3 again.
Group B: 4 and 5
These two are near enough the same size and shape that mixing them up is the single most common early error. They are also the pair that teaches you to stop judging by outline and start reading bands.
- 4 bands evenly. The dark bands are spread down the q arm at fairly regular intervals, like rungs on a ladder, with no single band shouting louder than the others.
- 5 is uneven. There is prominent dark banding in the proximal half of q, and then the distal q goes noticeably pale toward the terminus.
Say it to yourself as 4 is a ladder, 5 fades out. At lower banding levels that difference in the distal q is often the only thing you get, so learn it as your primary discriminator rather than your backup.
Group C: 6 to 12, and X
The crowd. Work through them in this order.
- 6 has a distinct pale window in the middle of the p arm, sitting between darker material above and below. Group C plus an obvious window in p is a 6 until proven otherwise.
- 7 reads top down: a pale cap at the end of p, then a dark band, and then a q arm with clear dark stripes and a lighter bottom.
- 8 also has a lighter p arm, but its tell is on q: a dark band close to the distal end, near the terminus rather than mid-arm.
- 9 is the lady in sunglasses and a dress. Two dark bands near the top read as the glasses, and the pale, variable block just under the centromere on q (9q12, another polymorphic region) makes the waist of the dress. Overall, 9 alternates strongly between light and dark.
- 10 has a band running across the p arm plus, at 400 to 550 bands, three prominent dark bands on q. Those three blocks are the standard trick for estimating resolution, which is why 10q is the chromosome the banding guide is built around.
- 11 is identified from the top: the distal p arm has a characteristic pale-then-dark pattern, and q carries two distinct dark bands with pale between them.
- 12 looks like 11's cousin and has a single dark band in proximal q sitting inside pale material. Push the resolution up and that one band resolves into separate bands, which is a useful self-check that your spread is better than you thought.
- X is the other lady in a dress, and this is the classic C-group confusion. Distinguish it from 9 by overall tone: the X is more uniformly grey-dark along its length, where 9 gives you high-contrast white and black. If the "dress" has strong black-and-white contrast, it is a 9. If it is muted and even, it is an X.
Two sanity checks for group C. First, count. You should end up with sixteen chromosomes in group C in a female (6 to 12 in pairs, plus two X) and fifteen in a male. Second, if two "9s" and no X turn up, you have almost certainly mislabelled the X.
Group D: 13, 14, 15
All three are acrocentric and all three look, at first glance, like the same chromosome. They are jack-o-lanterns, and the mouth is what differs.
- 13: eyes plus two big dark bands mid-q, thick and close together. The full mouth.
- 14: eyes plus a single strong dark band, straighter and flatter, with a second band splitting out below it.
- 15: eyes plus a small, faint mouth. The proximal dark band is there, but 15 is the quietest and shortest of the three.
Size helps as a tiebreaker (13 is largest, 15 smallest) but size alone is unreliable in a stretched spread. Read the mouth.
Group E: 16, 17, 18
- 16 is metacentric with a pale p arm and a striped q arm, which is where the bumblebee comes from: plain head, striped abdomen. It also carries a polymorphic heterochromatic block at 16q11.2, so a fat-looking proximal q is normal variation.
- 17 has a pale p arm and a pale proximal q, then darker material distally. The silhouette reads as a small figure in a summer dress.
- 18 is a teddy bear: a small p arm for the ears, then two dark bands on q making the head and body. It is also the chromosome the banding guide uses as a triage check, because if you cannot resolve 18's two q bands the cell is probably not worth analysing.
Group F: 19 and 20
Both metacentric, both short, both conspicuously pale, which is exactly why they are easy. Almost nothing else in the karyotype is this washed out.
- 19 is the palest chromosome in the spread, with a dark centromeric region and very little else.
- 20 looks like 19 but slightly darker, with a visible dark band on the p arm.
If you are staring at a small pale metacentric chromosome and cannot decide, it is a 19 or a 20, and that is already most of the answer.
Group G: 21 and 22
Both small acrocentrics. They are the bowling pins.
- 21 carries a strong dark band immediately below the centromere and tends to look stubby and broad.
- 22 has the same architecture with a paler proximal band and a whiter distal q.
Here is the fact the exam loves: 21 is smaller than 22. The numbering is historically wrong and was deliberately left alone, because by the time anyone confirmed it, "trisomy 21" had been in the literature for years and renaming it would have been worse than living with it. Size is therefore a legitimate discriminator in this pair, and it runs backwards from what the numbers suggest.
X and Y
- X is a group C submetacentric, medium sized, evenly and fairly darkly banded, with a prominent dark band in the mid q arm. Compare against a 9, not against your memory.
- Y is a group G acrocentric whose distal q is solidly dark (the Yq12 heterochromatin). It is the most length-variable chromosome in the genome, so a short Y or a long Y in an otherwise normal male is usually a polymorphism and not a finding. Unlike the other acrocentrics, the Y has no satellites.
Putting the karyogram together
Watching somebody sort is passive. Sorting is where the learning happens, and there is more than one way to do it.
In the walkthrough, the same unsorted spread gets rebuilt three ways, and they suit different people:
- Dragging. Slowest, and the best starting point, because your hand is committing to a position while your eye is still checking. Spatial memory is doing real work here.
- Pressing Enter. The tool places the next chromosome for you. Useful once you recognise chromosomes faster than you can drag them, and good for building speed.
- Typing the number. Fastest by a distance. You look, you decide, you press
4, and it lands. Twenty-three decisions in under a minute once you are fluent, which is exactly the repetition rate that turns recognition into reflex.
Start on dragging. Move to typing as soon as dragging starts to feel like the bottleneck. The goal is not to finish a karyogram; it is to make the decision this is a 14 happen without deliberation, and the number keys get you there fastest because they remove everything between deciding and confirming.
Orientation, once it is placed
A chromosome in the right box but the wrong way up is still wrong. Short arm goes up. In CytoSnap, F flips a chromosome, and dragging near the edge of a placed chromosome rotates it so you can straighten anything that came off the spread at an angle. Press the keyboard icon for the full command list, including auto-assign and magic arrange.
Getting orientation right is not cosmetic. The reason the convention exists is that the eye compares two homologues by scanning down from a fixed reference point. Flip one partner and you have destroyed the comparison the layout was built to make.
What trainees get wrong
| Mistake | Why it happens | The fix |
|---|---|---|
| Sorting by size alone | Size is the fastest signal, so it becomes the only one | Force yourself to name the centromere position out loud before you place |
| Calling an X a 9 | Both read as "lady in a dress" | Check contrast: 9 is high contrast, X is even and muted |
| Swapping 4 and 5 | Nearly identical size and shape | Read distal q: 4 stays banded, 5 fades pale |
| Treating 21 as bigger than 22 | The numbers imply it | 21 is smaller. Learn it as an exception |
| Reporting polymorphic heterochromatin | A big 1q12, 9q12, 16q11.2 or Yq12 block looks like extra material | These four regions are normal variants |
| Calling acrocentric p arms abnormal | Stalks and satellites vary a lot between people | Normal variation, not a finding |
| Sorting a spread that should have been rejected | Wanting to finish the cell you already started | If 18's q bands are unresolvable, pick another cell |
| Skipping the count | Sorting feels more productive than counting | Count first, every time. A 45 or 47 changes what you are looking for |
From a sorted karyogram to a reported result
Sorting one cell is a training exercise. A clinical result is a different animal, and it is worth knowing where the exercise sits in the real workflow.
For a routine constitutional blood study, the usual practice is to count around 20 metaphases, fully analyse at least 5 band by band, and karyotype 2 (one from each of two independent cultures where possible). Counts go up when mosaicism is a question, and cancer studies follow different protocols entirely. Your laboratory's SOP is the authority; these numbers are the common baseline behind it.
Counting, analysing and karyotyping are three different jobs. Counting is chromosome number only. Analysing is comparing homologues band by band at the microscope or on screen. Karyotyping is producing the sorted image. A normal count with no analysis is not a normal result, which is why the numbers above are stated separately.
Then you write it down, and ISCN is its own skill with its own rules. The ISCN nomenclature cheat sheet covers the grammar, and the top 25 chromosomal abnormalities covers what you will actually be writing most often.
A two-week practice plan that works
- Days 1 to 3: groups, not numbers. Sort into A through G only. Do not try to name individual chromosomes. You are training the size-and-centromere reflex, and it is the foundation everything else sits on.
- Days 4 to 6: the easy neighbourhoods. Add 1, 2, 3, 19, 20 and the acrocentrics. These are distinctive enough to build confidence fast.
- Days 7 to 9: group C. The hard one. Drill 6 to 12 plus X repeatedly. Expect to be bad at it for two days, then suddenly not.
- Days 10 to 12: the confusable pairs. 4 against 5, 9 against X, 21 against 22, 13 against 14 against 15. Do them as isolated pairs, not as whole spreads.
- Days 13 to 14: speed. Full spreads, typing the numbers, timed. Do not chase accuracy here. Chase fluency, and accuracy follows.
Ten minutes a day beats one two-hour session a week. This is motor learning as much as it is knowledge, and motor learning wants frequency.
FAQ
How long does it take to learn to karyotype? Recognising groups takes days. Naming most chromosomes reliably takes a few weeks of daily practice. Being fast and confident on poor-quality spreads takes months on a real bench, and that part cannot be rushed by any tool.
What is the easiest chromosome to identify? Chromosome 1 (largest, with its dark 1q12 block) and 19 (palest). Chromosome 3's pale centromeric sandwich is a close third.
Which chromosomes are hardest? Group C: 6 to 12 plus the X. Similar size, similar centromere position, seven of them plus a sex chromosome in the same size bracket.
Why is chromosome 21 smaller than chromosome 22? The original numbering was based on size and got this pair backwards. By the time it was confirmed, Down syndrome was already established in the literature as trisomy 21, so the numbering was kept deliberately.
Do I need a microscope to practise? No. Karyotyping practice is pattern recognition on images, and it transfers directly from screen to microscope. CytoSnap runs in a browser, and a free CruxSci account also lets you upload one metaphase of your own.
Is this on the ASCP CG exam? Yes, heavily, and much of it as image-based questions. Chromosome identification, banding landmarks, normal variants and ISCN description all appear. The ASCP CG study guide maps the full content outline.
Can our laboratory use this for training staff? Yes. CytoSnap accepts your own metaphase images, so new techs can train on the cases and the banding quality they will actually see. If you want to set this up across a team, get in touch.
Keep practising
Free, no account: CytoSnap Case 1 and the free 20-question CG assessment.
Free account: Case 2, plus one metaphase upload of your own.
CruxSci membership: every case, the Cutting Lab where you cut a raw spread into individual chromosomes yourself, unlimited uploads, the full ASCP CG course, and six computer-adaptive practice exams drawn from a continuously reviewed bank of over 1,000 questions.
Flow cytometry and molecular biology modules are next, and beta is open for both. If you work in either lane, we want your eyes on the material before it ships: CruxFlow beta for SCYM and flow, CruxMol beta for MB and molecular.
Keep reading
- How to Read Chromosome Banding (Without Losing Your Mind), including the 10q resolution trick
- ISCN Nomenclature Cheat Sheet
- Top 25 Chromosomal Abnormalities Every Technologist Should Know
- The History of Chromosome Banding, which explains why the stripes exist at all
- FISH Probe Types Explained, for when banding runs out of resolution
- Life as a Cytogenetic Technologist
- Which ASCP Exam Should You Take?
- Browse every CruxSci article