Q: Does this tool plot my family tree?
A: We can't gather enough data from DNA yet to be completely sure about anyone's ancestry. Think of SAPP as a modeling tool rather than a discovery tool - given what you know about the STRs, SNPs, and genealogies for your group it will help you plot the likely relationships between members.
Q: How do I know if the tree is right?
A: As you look at the trees produced by SAPP, look for inconsistencies. Do the SNP labels line up with the SNP test results? Do the patterns of STR mutations make sense, or does the same STR mutation show up repeatedly in close branches of the tree? Is the number of STR mutations in certain parts of the tree unreasonably large for the timeframes?
If you spot inconsistencies, the first likely reason is that the constraints on the tree are wrong. Check that the SNP and genealogy cross-reference matches the SNP and Genealogy inputs you meant to provide and correct anything that looks wrong. Check the SNP marker table to ensure that the kits were read in correctly. If everything looks right with the input then consider whether your assumptions about the group are correct. Are you sure those members are ALL descendants of that ancestor? Could one of the SNP results be incorrect?
You may end up with more than one possible tree as you change inputs and see the results. Hopefully that will help you prioritize what testing your members should take next in order to find out more about their shared heritage.
Q: Why does this program work best for R1b-L21?
A: SAPP has built-in a lot of data about the L21 haplogroup and its sub-clades. Besides the L21 modal haplotype, it also knows the SNP sub-tree under L21 (based on this tree from the R1b-L21 Project). It also contains the allele frequency values specifically for L21. This data is not applicable to other haplogroups and may cause SAPP to make incorrect conclusions for haplogroups outside the L21 SNP tree unless the correct data (modals, etc) is entered.
Apart from that data SAPP also knows the STR mutation rates for STRs up to the 111 tested by FTDNA, but those are universal across all haplogroups and not specific to L21.
Q: Why does SAPP switch to showing trees in text form for larger numbers of kits?
A: Because the picture size needed to show large trees gets too large to manage. With 100 kits of input, the resulting picture can be over 10,000 pixels on each side! So to allow the tool to show results for larger numbers of kits, SAPP offers the text-based tree structure as an easier option to handle for showing the relationships, STR mutations, and TMRCA data.
Q: Why do the TMRCA estimates sometimes go DOWN farther up the tree?
A: The TMRCA estimates are based on the genetic distances between the kits in the node's branches. Those genetic distances are not proportionally spread across time and so the resulting TMRCAs can be higher or lower at different points on the tree. Also, the program averages across the branches so fewer or more STR mutations can sometimes affect the results.
The TMRCA estimates show the most likely estimate first, with the 5%-95% ranges shown in brackets. Normally no matter how the most likely estimates change as you go back in time, the ranges will still get older as you move further up the tree. In rare cases, however, the entire TMRCA estimate may look out of place. This may indicate that fewer or more STR mutations than normal occurred between those branches, or it may indicate that SAPP was wrong in how it placed those branches on the tree. You'll need to look at the STR mutations carefully in that area of the tree to see if they make sense.
In the end, remember that TMRCA calculations based on STRs are only an approximation. As you find out more about the timeframes your ancestors lived in and the projected ages of the SNPs used in your tree, use those age estimates to calibrate the TMRCAs that SAPP reports for the various nodes.