INTRODUCTION 

Most of the relationships we sought to prove are too many generations ago to give reliable DNA matches through conventional means. The potential link between for example the English Lousadas and their 10th and 11th cousins among the USA Lousadas goes back 9 generations to different siblings of Amador de Lousada of Vinhais. Even 4th cousins do not show up well, for I have only one (25cM) segment in common with my 4th cousin Jeremy Lousada, and only 3 3cM matches with my 6th cousin John Griffiths. From the outset, we were well aware of the poor match-prediction powers of small (3cM) segment matches. Even at 7cM there is significant uncertainty as to their reliability, and only going to 15M gives this. Therefore we sought to rely on maximising the number of matches and then using statistical and related techniques to discern patterns lying within our datasets.  

As Qmatch was recommended to us for small matches by GEDmatch, and retaining our desire to accumulate many matches, we set about using Qmatch (set at 3cM, P=3) to compile all 2255 segment matches - mostly off-target matches and false positives of course -  from our 13 relative sample - see here the 1963 matches found before ELL was added to the sample of relatives. We found 46 RSBCs, 25 lefthand and 21 righthand, and most were new to us. We spent much time looking at RSBCs, how frequently they occur across chromosomes and in the match-rich areas on chromosomes. But they proved quite difficult to work with, and were relatively unproductive.

 We then moved from RSBCs to ASBs; but they proved misleading and just as we prematurely claimed in  'Fun with Autosomal DNA', we again thought that we may have established a genetic link between the USA Lousadas, the English Lousadas, the Barrows, Scott's wife (and hence the Fischls) and Randy's parents. The problem with ASBs was that instead of an ASB being associated with a unique crossover, in fact many unrelated crossovers can all report to the same ASB (in fact the same applies to RSBCs as well). This is because a crossover is only measured as lying between 2 particular SNPs - these SNPs being the first non-matching SNP beyond the crossover at one end of a segment and last SNP within the segment at the other end of the segment. Typically there are 5000 base-pair positions between each SNP so the potential for unrelated and (from the family viewpoint) spurious crossovers at the same ASB is large. That is, ASBs are not the amazingly precise way of penetrating the fog of unreliable small matches that was hoped.

From Chart 1 we can see the futility of the 3cM P=3 results. For at this setting the randoms show more segment-matches than the relatives - 2261 to 2255 - that is, family does not show a signal! While the RSBC totals indicate that a family signal may exist, the ASB numbers seem odd (and perhaps can only be explained by a combination of spurious crossovers and the much greater possible ancestor pool where there is no genealogy).

 

 CHART 1- COMPARISON DATA RELATIVES AND RANDOMS

 

AT LAST THE RELATIVES STAND OUT 

From this impasse it was timely advice from GEDmatch in February 2026 which assisted us. Thus, we came to use Qmatch to look again at all 3cM segment-matches, not at P=3 but at P=7. These conditions give reasonable quality segment-matches, almost as good as 7cM matches from some other providers. And most importantly Chart 1 shows that at 3cM P=7, the relatives show 560 segment-matches which is 187 more than the randoms' 373 segment-matches, the strongest such family signal we have seen in all our comparisons of relatives with randoms. Chart 2 shows the segment-match numbers for relative pairs. For example, among the 14 B/Je segment-matches is one of 5cM on Cr10 which is presumably their Ancestry.com 7cM match. JG misses matching (shown in grey) 2 of the Barrow-Lousadas (A and J) whose specific ancestry evidently differs from that of Ju.  

CHART 2 - RELATIVES SEGMENT-MATCHES

 

COMPARISON OF RELATIVES AND RANDOMS 

JG's 2 misses in Chart 2 contrast with 9 misses by the randoms in Chart 4. Further, we see that the randoms have 4 people in less than 4% of segment-matches compared with 1 relative (J, who despite proven close Lousada genetic and genealogical links, has anomalously low segment-match numbers), and 4 people in more than 10% of segment-matches (compared with 1 relative). That is, as expected, family connections despite some stochastic phenomena also being present, are tighter. The presence of 5cM and 7cm segment-matches in Chart 4 is indicated by the same colours as in Chart 2, but also with pale blue showing the pairs with 8 segment-matches (for the reason explained below). Our first task is to assess how well our grouping of relatives (indicated by lines across the above table) compares with random. We can see in Chart 3 that the intrabranch matching within each of the 3 groupings comfortably exceeds that of the random set, and supporting this we note that a triangulation is shown by the 1st and 4th cousins A, Ju and J at Cr16 (3219600 - 6259081) within the Barrow-Lousadas. Other triangulations are discussed below.

CHART 3 - INTRABRANCH MATCHING

 

ANALYSIS OF RANDOMS 

The random set could contain its own (but unexpected/unknown) internal family links which must be allowed for when we consider the real segment-match count and the false (off-target) segment-match count. Chart 4 shows the random set does indeed have structure with several obvious groupings. The central grouping seems to reflect several (perhaps 3) Ashkenasi families variously linked perhaps 4-8 generations ago. A second grouping is more distantly related perhaps with more diverse ancestry. The outer grouping is only remotely related. From this, we can make an estimate of the number of real segment-matches among the 373 in the random set - simply by augmenting the central group's 154 segment-matches by the number of segment-matches emerging from the surrounding group. For this we need a cut-off number N to define which pairs to include. With N=8 we obtain 4 pairs showing 8 or more segment-matches to arrive at an estimate of 38 real segment matches, which we add to 154 making 192, hence 181 false matches. The 3 8-segment matches are shown in blue while the 12-segment match is left green. 

 

CHART 4 - RANDOMS SEGMENT-MATCHES

 

IMPLICATION FOR RELATIVES 

Discovering the correct number of segments (N) to use in deducing real and false relative segment-match numbers in practice proceeded iteratively, but we can summarise the logic here by which the selection of N=8 was made. In this, as was seen, we were mindful that false yellows emerge from the triangulation discussion below. Once we made the choice that N=8, we subtracted 181 from the total (560) relatives segment-matches to arrive at 379 real family segment-matches; but more importantly we could deduce what this means for proven matching of each pair of relatives. For now, having a probability estimate covering the false segment-matches (namely 32% or 181/560), a pair with 8 segment-matches is tantamount to a proven match since 0.32**8 = 0.0001 or 0.01%.

CHART 5  - RELATIVES REAL MATCHES

Accordingly, we can now present in Chart 5 our estimate of real relative matching. Here the yellow, green and blue colours show matches between relatives. The blue entries show 8-segment matches not already included as yellows or large (8 or more segment-matches) greens. The yellows have been reduced by the known false TP 7cM match (see triangulation discussion below).

 

INTERBRANCH CONNECTION 

There are 4 branches of relatives - English Lousada (MD), Barrow-Lousada (Ju, A, J), Barrow (E, JG, RM, RM, SW), and US Lousada (ELL, B, Je, TP). We extract from Chart 5 the number of matching relative-pairs for each of the 6 interbranch intersections, and express this as a percentage of total possible relative-pairs - see Chart 6. From this we can make some observations:

 

CHART 6 - EXTENT OF INTERBRANCH CONNECTIONS

 

TRIANGULATIONS 

Triangulations show 2 or 3 branch matching - Ju/RM/Je at Cr2 (217m - 220m), B/Je/E at Cr5 (79m - 81m), J/TP/E at Cr10 (116m - 119m), RM/E/A at Cr17 (31978888 - 33622774) and RM/B/Je at Cr22 (25640628 - 26225384). But at the Cr2 site just noted the triangulations Ju/Je/TP and B/Je/TP are likely to be false (containing at least 2 non-Lousada matches as Ju does not match B here). And of the 5 remaining triangulations, only the 2nd triangulation (B/Je/E) is entirely composed of 8-segment matches and thus appears real along with the obvious Ju/J/A triangulation noted above. The absence of a triangulation can also be informative - on Cr18 (6.6m - 8.1m) each of the 1st cousins Ju and A has a good match with TP but don't match each other here which shows that at least one of the 2 7cM matches is in fact false - probably A/TP. This of course shows that false matches can occur in the low-match yellow pairs, which we allowed for above.

 

CONCLUSION 

We may feel confident that despite the small segment sizes and remote ancestral connection, there is considerable interbranch matching which supports our genealogy (in which the various branches originate in the ancestral family of Amador de Lousada).

 

FOOTNOTE 

Finally, out of respect to those whose kit numbers went into generating the random sample, we comment further on Chart 4. This shows perhaps 3 family linkages around 4-8 generations ago (A/H/J/MW/DG) with a secondary group (N/S/M/P), together with an outlying group (C, MB, MMC, KB). The central 'family' group of 5 contains kits contributed from John Griffiths (A, H, J) and Julian Land (MW, DG). The secondary group of 4 consists of N contributed by John Griffiths and S, M, P contributed by Julian Land. The remotely-connected group of 4 consists of C contributed by John Griffiths and MB, MMC and KB contributed by Julian Land. The set of 13 randoms produced only 2 triangulations, on Cr1 at 158m - 160m (H, S, J) and on Cr8 at 11.3m - 12.6m (S, P, M). These triangulations mean little in this context but we note that the second has one 8-segment component S/P.