Background:
For both Color by and Filter for SAMtags, users will select/enter a SAMtag, such as CB,CR,UB,MI, and the corresponding values, eg. TGTACTCAGA, that they are interested in seeing or coloring. At any instance, users can only select on SAMtag but they can add multiple values of interest. When the users click OK, this information is then packaged and handed over to SAMtagsFilter or SAMtagsColor to make corresponding changes in the track.

Each read in the track is compared with the user values one by one and then its filtered/not filtered or colored/not colored. For SAMtagsFilter and SAMtagsColor, this conditional step happens in particular function and these functions are called for each and every read alignment:

• For SAMtagsFilter (extends SymmetryFilter), the function is filterSymmetry(BioSeq seq, SeqSymmetry sym) and,
• For SAMtagsColor(extends ColorProvidor), the function is getColor(SeqSymmetry sym)
  Every time the function is called read alignment information is passed.

To begin the investigation, I suggest starting from the algorithms in these functions. For single-cell data, the number of reads in a given track could be a lot and processing through each of them could take up some computing power. There was an earlier issue where the filterSymmetry function was called twice for each read which was fixed in an earlier ticket. Considering the volume of data that needs to be handled, it is possible even a small change in the efficiency of the algorithm could have a huge difference.

Files needed are located in Google Drive.

Steps to reproduce the issue:

1. Load the data
   1. Load the hg38 human genome
   2. Go to chr16:28,931,601-28,939,711
   3. Add the 10k_PBMC_chr16.bam file
      1. Download the file from Google Drive
   4. Click Load Data
2. Install the Monster Alignment Filter
   1. Go to Plug-ins, Launch App Manager
   2. Select the Monster Alignment Filter
   3. Click Install
   4. Close the App Manager
   5. Right-click the track label, select Filter...
   6. Click Add > Reads without monster gaps, set Max intron size to 4000, click OK
3. Color reads by cluster
   1. Right-click the track label, select Color by...
   2. Select SAMtags > Edit Tags and Color
   3. Click Import... and use the file Graph-Based_10K_PBMC.csv and click Open
   4. Select SAMtag: CB
   5. Click Save and Apply > OK
   6. Wait for the colors to change
4. Duplicate the BAM file track
   1. Right-click the track label, select Track Operations > Copy
5. Color the copied track by cluster
   1. Right-click the new Copied track label, select Color by...
   2. Select SAMtags > Edit Tags and Color
   3. Click Import... and use the file Graph-Based_10K_PBMC.csv and click Open
      1. Note: the color values may already be set, it is not necessary to import the file again
   4. Select SAMtag: CB
   5. Click Save and Apply > OK
   6. Wait for the colors to change
6. Filter the original BAM file by Cluster
   1. Right-click the original BAM track label, select Filter...
   2. Click Add > SAMtags
   3. Set SAMtag to CB
   4. comparator =
   5. For the value copy/paste the contents of 10k_PBMC_cluster10.txt
   6. Click OK
   7. The reads should all be brownish/red
7. Filter the Copied BAM file by Cluster
   1. Right-click the Copied track label, select Filter...
   2. Click Add > SAMtags
   3. Set SAMtag to CB
   4. comparator =
   5. For the value copy/paste the contents of 10k_PBMC_cluster11.txt
   6. Click OK
   7. The reads should all be green
8. Load data for a different gene
   1. Go to chr16:28,950,223-28,967,238
   2. Click Load Data
   3. Wait for the data to load and colors to change
      1. You may need to click Load Data again to get the data to load for the Copied track

Pulled the latest commits in IGBF-4295 and created a new branch IGBF-4344 from IGBF-4295.

Changes :
Created a single HashMap that stores tag values and colors when the color file is loaded, and this HashMap is rebuilt only when a tag value or color value changes. This improves performance because it is no longer rebuilt every time the getColor method is called.

Commit Id: https://bitbucket.org/lorainelab_udaya/integrated-genome-browser/commits/572202179e1a05e50346f9c44c01f4eb33c5a3b4

Cc : Karthik Raveendran, Dr.Nowlan Freese

Tested Udaya's branch on Mac.

Following the testing instructions above, the speed of loading data is vastly improved. IGB no longer hangs when loading data. With sequence and data loaded for the view chr16:28,799,070-29,019,552 the memory usage is equivalent to the same data loaded in IGB 10.1.0 release (around 950Mb).

For the PR, I believe Karthik Raveendran will include this commit on top of his most recent changes for IGBF-4295.

Made changes to improve the performance of SAMtools Tag Color Mapping table.

Commit Id : https://bitbucket.org/lorainelab_udaya/integrated-genome-browser/commits/6af2c02b4df51b853ca1af9209a8ee4110449976

Cc: Dr.Nowlan Freese, Karthik Raveendran

Fixed the issue caused when samtag value is selected.
Commit Id : https://bitbucket.org/lorainelab_udaya/integrated-genome-browser/commits/f9fee8c4876d5552deb9eef095a94bbd70e7ef1e

Cc : Dr.Nowlan Freese, Karthik Raveendran

Tested Udaya's branch on Mac. Working as expected.

Karthik Raveendran - please add Udaya's commits to your 4295 branch.

Udaya's commits are included within Karthik's pull request: https://bitbucket.org/lorainelab/integrated-genome-browser/pull-requests/1088

PR is merged.

Closing ticket.