Sunday, January 22, 2017

GMP Performance

I know performance is a point of frustration with Genome Mate Pro (GMP) for some.  I can see a significant difference when loading data into my husband’s computer versus mine.
There is no doubt that GMP is much slower importing data than the old app. This is because the old app kept all data in memory and saved to disk every 5 minutes. Memory access is much faster than disk access but due to a Silverlight constraint, the in-memory database was limited in size.
The new app is virtually unlimited in size as it keeps the data in an SQLite database on disk.  However, the time it takes to load is a function of disk access speed and this varies from user to user.  I have been told that a solid state drive (SSD) is very efficient.
One way to improve performance without replacing your hardware is to set the import criteria in Options as highlighted below.  This will reduce the number of writes to the log file.
Starting with the 2016r01 release, it will be possible to enable in-memory journaling.  What this does is to write the SQLite journal file to memory rather than disk.  This is not recommended because there is a risk that if GMP is shutdown abnormally, the database may become corrupted and unusable.  If you do choose to use this option, BACKUP (File -> Backup) first.
~ Becky ~
P.S.  Tested loading data on my husband’s slow computer with these two options and the import time was amazing (197k records in less than 30 mins).

Import Issues?


Having issues importing data?  Data not there or rejected?  Formats occasionally change but most of the time the issue can be resolved by following these steps:
  1. Verify that the data is not being hidden because of a filter selection on the Chromosome Browser page.  These should be set as above image to view all segments. 
  2. Verify that the correct import option was selected.  There are many different import sources and sometimes multiple providers of the same data so ensure the import option reflects both the correct source and provider.
  3. Verify that the data in question is contained in the import file and if using copy/paste that the data is copied from the Chrome browser.
  4. Verify that the keys (name, kit, etc.) in the Profile record are entered correctly with the names and keys exactly as they are listed in the source data file.  Otherwise, the app will not be able to identify the correct profile.
  5. Review the error generated in the log file.  Jim Sipe’s User Guide contains a list of error messages and possible resolutions. No log file? Verify that the log file is not open in another program such as Excel.
  6. Verify that the settings are correct in Options -> App Settings
When the steps above have been verified and the issue is still not resolved then post the following screen shots on Facebook.  Be sure to include the version of the app and type of computer being utilized.

  • Snip of the profile page in question
  • Snip of the error in the log file
  • Snip of the App Settings in Options
  • Snip of GMP page (Relative, Chr Browser) showing problem

Please note that in options you can select Privatize the Display for Sharing before taking a screen shot.

If we cannot identify the issue from the screen shots, you may be asked to email your import file to me for further analysis on my test database.


~ Becky ~

Triangulation and ICW in Genome Mate Pro

The purpose of this post is to explain how triangulation and “in common with” (ICW) work in Genome Mate Pro.
Chromosome Browser
From the Chromosome Browser page, right click on a Relative to see what segments triangulate with that person and the profile person.
The first thing to note is that the app only displays what has been imported. It has no means to automatically identify either of these so if it is missing in the display then look to the import log to see if there is an error message associated with the import.
The second thing to note is that triangulation and ICW data are only retained for relatives found in the database. This is different from the old version of the app and it is important to load relative data before loading chromosome data.
Triangulation
In Genome Mate Pro, triangulation is based on two people being related to the profile person on the same chromosome segment so triangulation data has a chromosome number, base pair start and end points associated with it.
For example, profile person P matches relative A and matches relative B on the same segment (P==A, P==B). If relative A also matches relative B on the same segment (A==B) then P, A and B are triangulated. Otherwise, either A or B is likely related to profile person P on their maternal side while the other is related on their paternal side.
Data is loaded via various chromosome browser files such as the 529andYou and FTDNA Chromosome browser .csv files as well as specific triangulation imports such as GedMatch’s Tier 1 Triangulation.  Triangulations loaded from chromosome browser data have the import limitations listed on the Options page imposed on them as well.
Some caution should be exercised when a displayed triangulation is based on a very small segment overlap between the Relatives A and B as the endpoints from different sources tend to break at different intervals and endogamy can also cause havoc with the results.
In Common With (ICW)
ICW relatives do not have the chromosome, start point and end point data associated between them (typically from FTDNA) and as a result, the profile person P can have an ancestor in common with both relatives A and B but even if A and B have an ancestor in common, it is not necessarily the same one that is in common with profile person P.
Genome Mate Pro will display the people who are in common with the selected relative on the same segment.
Relative Page
Triangulation and ICW data are displayed on the Relatives page for the segment selected in the DNA Match Segments table on the left.  Since a Relative that triangulates with the selected person on that segment is already considered ICW, there may not be an ICW displayed in both the Triangulation and the ICW lists.
In summary, triangulation data has a segment specific information while ICW just says two people are related but gives no information on what segment that relationship occurs.
~ Becky ~