We’re not geneticists or conservation scientists. Everything below is what we’ve learnt working through this ourselves — reading, asking questions, calculating COI’s, speaking with our genetic testing company, building our own matrices, learning new things and tweaking and incorporating them. If you’re new to this, and want to keep flock conservation in mind, we think this is the plain paddock workable version: no assumed background, but nothing dumbed down either, and hope it an be useful. Even in small flocks, this information about your flock can be fascinating to know, you will see the result of many generations playing out in your own paddock, and it can help with excellent breeding choices. We wish we had known all this at the start of our Damara journey – and hope it can help someone else. If anyone who has undertaken testing or tracks pedigrees and breeding populations of their own reads this and wants to sharpen or correct a point, we’d genuinely like to hear it.

Kinship Coefficient

Before COI, it’s worth starting one step earlier, with the number COI is actually built from: the kinship coefficient. Take any two animals — just any two sheep in the flock — and the kinship coefficient answers the question: how likely is it that a randomly picked gene copy from one animal is identical by descent to a randomly picked gene copy from the other? ” – with identical by descent just meaning both copies trace back to the same shared ancestor. Kinship coefficients can come from pedigree, telling us what the relationship is expected to be, or from genomics, telling us what the relationship actually is. It is a number than describes the relationship between the pair, between brothers, sisters, cousins or just any two random sheep. A kinship coefficient of zero means the two animals share no common ancestry at all. A kinship coefficient of 0.25 means something like a parent and its own offspring, or two full siblings. The kinship coefficient tells you how related two animals are to each other.

Coefficient of Inbreeding – COI

COI stands for Coefficient of Inbreeding. It is a number attached to one animal, and it answers the question: what is the probability (pedigree COI) — or what is the actual likelihood (genomic COI) — that this animal inherited two identical copies of the same gene, one from its sire’s side and one from its dam’s side, both traceable back to the same ancestor? If the sire and dam share no ancestors at all, COI is zero. The more ancestors they share, and the closer those shared ancestors sit in the pedigree to the animal in question, the higher the COI climbs. Pedigree COI is calculated with Wright’s path-counting formula, adding across every common ancestor shared by the sire and dam, with each generation of distance reducing by half that ancestor’s contribution.  Genomic COI is calculated from the animal’s DNA, using thousands of genetic markers to measure how much of its genome is likely identical by descent. Rather than estimating from the pedigree, it measures the actual genetic outcome.

A COI of 0.10 means a 10% chance that any given gene in that animal exists as two identical copies due to shared ancestry, rather than as two independently inherited versions. It matters because this is where inbreeding depression comes from: more and more of the genome pairs up identically, recessive defects get more chances to be expressed, and the animal – and breed – loses the vigor that comes from carrying two different versions of a gene at each position.

Kinship Coefficient and COI – not the same thing

Because the two ideas share the exact same arithmetic, (kinship coefficient measuring the relationship between two animals — the parents, and COI measuring something that happens in a third animal) its easy to think they are the same thing – but they are not. One describes a pair, the other describes an individual. But when the two animals in question are that offspring’s actual parents, the number itself doesn’t change between the two calculations — the lambs expected COI is equal to the kinship coefficient of its parents. So the reason the numbers match isn’t coincidence, but it also isn’t because they are the same measurement – it’s because the offspring’s inbreeding is just the parents’ relatedness, carried forward one generation.

Kinship Matrix

A kinship matrix takes the kinship coefficient and applies it to an entire flock at once. Instead of one number for one animal, you get a full table: every animal compared against every other animal, with a kinship coefficient in each cell. Full siblings, half siblings, cousins, complete strangers genetically — all quantified, all sitting in the same table. You can build a kinship matrix two ways. From pedigree alone, using recorded ancestry and standard inheritance rules — this gives you the expected relatedness – or from genomic data, using actual DNA markers to measure what was actually inherited — this gives you the actual inherited relatedness. Those two versions can diverge.

In practice, you don’t really “read” a kinship matrix top to bottom — you use it to answer specific questions as they come up. Thinking about a mating? Look up that sire and dam’s kinship coefficient, and you already know roughly how inbred any resulting lamb would be, before you’ve committed to anything. Want to check the whole flock for relationships that pedigree might be underselling? Scan the matrix for pairs that are closer than expected. It’s also what mean kinship is built from — averaging one animal’s numbers against everyone else’s tells you whether that animal is genetically rare and worth leaning on, or already well represented in the flock. The nice thing is it’s all the same table. Build it once, and a single pairing check, a flock-wide risk scan, and a ranking of your most valuable animals all come from the same source, not a separate calculation each time. Below is our current KM, 93 pairwise relationships – a fast reference for understanding each relationship in the flock.

Mean Kinship

Mean kinship (MK) answers two questions. Firstly, “how much of this animal’s genetic material is already circulating in the rest of the flock?” It is calculated from the kinship matrix, with mean kinship reducing the matrix to either one number per animal: the average of all its pairwise relationships against every other individual. A low individual MK means an animal’s genetics are underrepresented in the flock, and a high MK means they’re already well spread through it. This holds true no matter how distant any one pairwise relationship looks, and regardless of the animal’s own COI — because MK and COI are measuring different things entirely. COI looks backward at an animals own two parents and traces the shared ancestry between them; it tells you whether that particular mating risks inbreeding depression. MK looks outward, at how related that animal is to everyone currently in the flock. Two animals can have identical, very low COI — neither inbred by parentage — and still carry very different MK: one might be drawing on a rare founder line barely present elsewhere, the other on a line already used heavily across dozens of other animals. Same COI, very different genetic footprint. That’s what makes MK catch things a single pairwise relationship or COI misses entirely. Two animals can share no recent common ancestor and look like a clean outcross by COI or pair relatedness, and still both carry high MK, because each is drawing independently from the same narrow founder base further back. The reverse also happens: a sheep with few recorded descendants can carry a low MK and represent a genuinely rare genetic contribution, one that pedigree structure alone won’t reveal, as it’s the result of many small pairwise relationships across a large share of the flock.

MK secondly answers the question “How related is my whole flock to each other”? It is the average of all the distinct KM coefficients (including self kinship). It matters because MK really only ever climbs in a closed population – and tends to climb in a limited or unmanaged population. If diversity or sustainabilituy is your goal – your MK should be stable or preferably decline, and is arguably a much stronger strategy than simply calculating COI. The trap with COI-only breeding is that it’s easy to keep pairing your most common, best-connected animals together — the COI always looks fine — while your rarest bloodlines sit unused because there’s no low-COI match readily available for them. Every mating looks safe on paper, but generation by generation, the rare lines quietly get diluted out, and diversity is lost anyway. Prioritising low mean kinship flips that: it pushes us to actively breed from the animals whose genetics are underrepresented, keeping more founder lines alive in the population — even if it’s not always the mathematically “safest” pairing on a single-litter basis. Together … COI asks “is this mating safe?” and MK asks “is this the mating that does the most good for the flock’s long-term diversity?” We use both — but mean kinship is the one steering our flocks preservation picture.