Determining Matches after DNA testing is usually the first-order start at analyzing test results. In most cases, the matches are determined by the test company or the site that the RAW data results from testing are up loaded too. Determining a match is heavily dependent on the test performed. See the genetic genealogy testing tutorial for some further details.
Many test companies have a cut-off after which they will not report a tester in a match list (that is, as a match). A 7 cM minimum largest matching segment length or 10 cM of total matching segments sum are commonly used values. They may also use different criteria for determining whether a matching segment is included in the sum or in calculating the largest matching segment. Often matching segments of less than 500 SNPs or 7 cM in length may be considered as too unreliable and not used. But some test companies drop to as low as 200 cM and 1 cM when including matching segments in their sum. So further care must be taken when interpreting match list results with some test companies versus others. And can explain some of why two test kits will appear with different match values with two different test companies.
Some test companies include the xDNA matching segments in the summary totals given in a match list; others do not. The cut-off criteria for xDNA matching segment inclusion may be different. As well as whether to even include those who only have an xDNA match in the match list (that is, no autosomal matching segments).
Clustering is a common tool exploding in use starting in 2018. Variations exist but most are based on looking for shared match groups or clusters. So basically looking at all your shared match lists with all your matches and seeing patterns of comman entries between them. More rarely the tool may use matching segment overlaps for this process;which can be a more assured answer of a likely relationship but is often harder to find occurring among your (more distant) matches. Visual Phasing and Chromosome Painting were the ones before; with the former really being a form of segment clustering and Chromosome Painting.
Third Party Tools exist to manage Matches and Match Lists across the various test platforms and companies. GenomeMate Pro is one of the more popular desktop management tools and uses both the PedigreeThief Chrome extension and DNAGEDCom desktop app to gather the data from the various test companies. They gather the data and put it into a form readable, in a common way, by GenomeMate Pro. Genetic.Family, the next generation tool from the DNAGEDCom folks, is looking to support doing this cross-company analysis in a cloud-based (online server) instead of the desktop application like GenomeMate Pro. It is only with these larger, DB type tools that one can possibly begin to manage the thousands of matches obtained from today's larger databases (and due to looking deeper into the smaller total matching segments amounts below 40 cM / 0.5%).
Ultimately, having the list of actual matching segments used to create the summary numbers is the best. Especially if trying to compare across test company platforms. Then results can be normalized and compared more accurately. Or, more importantly, segment triangulation can occur to more definitively determine if there is a real, common ancestor between the matches.
Unlike autosomal DNA which quickly leads to no matching after 200 years or 6 generations to the common ancestor, match list membership here can typically exist for those with common ancestors outside the genealogical time frame (or roughly the last 500 years). Sometimes match list members may only share a common relative a thousand to ten thousand years ago.
For yDNA STR testing, a match list is commonly formed from those closest in genetic distance to the tester; possibly using a subset of the STR markers compared.
yDNA and mtDNA SNP test results are analyzed and placed in a phylogenetic tree of haplogroups. If a match list is generated, it is usually of those who are placed in the same or nearby branches on the tree. Depending on how deep the leaves are on the tree, those matches may not share a very recent common ancestor. "Deep" in this context means more recently formed in the past.
Deep WGS testing is beginning to find changes that always track within the last 100-500 years and thus allow for more discerning match lists and matches. Even causing some STR marker values discovered to be included as a phylogenetic tree branch designation (that is, defining a haplogroup). And removing some of the perceived uncertainties of TMRCA being observed with yDNA STR testing alone.
atxDNA matches
The matches term is mostly applied to autosomal test analysis as that is where matches are presented in a match list form; and sometimes that is all that is presented from those test results. (Well, ethnicity or Ad-mixture also but still a developing science only beginning to help with understanding genealogy at the present time.) The match list is created from those other testers who match some portion of the primary testers DNA. Often rank-ordered by the "strength" of the match as measured in centiMorgans. For autosomal testing, the match is determined by first finding matching segments between the primary tester and each other tester in the database. Then summing the matching segment lengths to create a total matching segments sum (in centiMorgans. Or possibly given as a percentage using ~72 cM as 1%. Often the test company will also report the number of matching segments and possibly the largest matching segment length (in cM). All three values are clues as to how strong the match is and thus how close a relation between the two testers.Many test companies have a cut-off after which they will not report a tester in a match list (that is, as a match). A 7 cM minimum largest matching segment length or 10 cM of total matching segments sum are commonly used values. They may also use different criteria for determining whether a matching segment is included in the sum or in calculating the largest matching segment. Often matching segments of less than 500 SNPs or 7 cM in length may be considered as too unreliable and not used. But some test companies drop to as low as 200 cM and 1 cM when including matching segments in their sum. So further care must be taken when interpreting match list results with some test companies versus others. And can explain some of why two test kits will appear with different match values with two different test companies.
Some test companies include the xDNA matching segments in the summary totals given in a match list; others do not. The cut-off criteria for xDNA matching segment inclusion may be different. As well as whether to even include those who only have an xDNA match in the match list (that is, no autosomal matching segments).
Clustering is a common tool exploding in use starting in 2018. Variations exist but most are based on looking for shared match groups or clusters. So basically looking at all your shared match lists with all your matches and seeing patterns of comman entries between them. More rarely the tool may use matching segment overlaps for this process;which can be a more assured answer of a likely relationship but is often harder to find occurring among your (more distant) matches. Visual Phasing and Chromosome Painting were the ones before; with the former really being a form of segment clustering and Chromosome Painting.
Third Party Tools exist to manage Matches and Match Lists across the various test platforms and companies. GenomeMate Pro is one of the more popular desktop management tools and uses both the PedigreeThief Chrome extension and DNAGEDCom desktop app to gather the data from the various test companies. They gather the data and put it into a form readable, in a common way, by GenomeMate Pro. Genetic.Family, the next generation tool from the DNAGEDCom folks, is looking to support doing this cross-company analysis in a cloud-based (online server) instead of the desktop application like GenomeMate Pro. It is only with these larger, DB type tools that one can possibly begin to manage the thousands of matches obtained from today's larger databases (and due to looking deeper into the smaller total matching segments amounts below 40 cM / 0.5%).
Ultimately, having the list of actual matching segments used to create the summary numbers is the best. Especially if trying to compare across test company platforms. Then results can be normalized and compared more accurately. Or, more importantly, segment triangulation can occur to more definitively determine if there is a real, common ancestor between the matches.
yDNA and mtDNA matches
yDNA and mtDNA testing can result in a match list of a different sort. This because matching segments are not determined to compare individual testers. Instead, actual marker values are compared directly.Unlike autosomal DNA which quickly leads to no matching after 200 years or 6 generations to the common ancestor, match list membership here can typically exist for those with common ancestors outside the genealogical time frame (or roughly the last 500 years). Sometimes match list members may only share a common relative a thousand to ten thousand years ago.
For yDNA STR testing, a match list is commonly formed from those closest in genetic distance to the tester; possibly using a subset of the STR markers compared.
yDNA and mtDNA SNP test results are analyzed and placed in a phylogenetic tree of haplogroups. If a match list is generated, it is usually of those who are placed in the same or nearby branches on the tree. Depending on how deep the leaves are on the tree, those matches may not share a very recent common ancestor. "Deep" in this context means more recently formed in the past.
Deep WGS testing is beginning to find changes that always track within the last 100-500 years and thus allow for more discerning match lists and matches. Even causing some STR marker values discovered to be included as a phylogenetic tree branch designation (that is, defining a haplogroup). And removing some of the perceived uncertainties of TMRCA being observed with yDNA STR testing alone.