Bug Fixes and Tool Updates
- Issue: Displaying only the top HSP alignment in the "Full Length Sequence Search Alignment Details" table when a query hits multiple locations in a single subject sequence.
- Solution (January 26, 2023): Updated details table so that the corresponding alignment is shown for instances of hits to multiple HSPs
- Issue: Formatting issues in the query sequence
- Solution (January 23, 2020): Create a pre-screen of the query sequence to check for possible formatting issues and/or for the presence of invalid aa characters.
- characters that the tool considered invalid are O (the letter), spaces, numbers, and special characters (! @ # $ % ^ & * ( ) - + \ } [ ] , : ' " ? . / \ ; _)
- the submit button is gray(disabled) initially and if input sequence fails validation (also shows error message "Invalid input sequence."); will only turn green (enabled) if it passes the sequence screening validation
- Issue: No Response output for all searches (full-length, 80-mer, and 8-mer)
- Solution (January 23, 2020): change "no response" output to "no alignment with E-value < xx"
(xx being whatever E-value the user has selected − "default" or other value from the drop down menu - as part of the input);
if it does not return any result other than the option above, it outputs "No alignment found"
- Issue: With the 80 mer searches output only having a single window 1:80 despite many of my proteins being longer than 80 amino acids
- Solution (January 23, 2020): adopt the "table view" consistently in the output display, showing all individual sliding windows in column 1 and the corresponding result for each row in column 2,
independently of the results being “none” (no alignment), "n alignments" (where "n" is any number of alignments), or a mix of both. If no alignments are found, "None" will be displayed in column 2
- Issue: 80-mer search showed results that extend beyond the length of the sequence
- Solution (January 23, 2020): Verify that the second number in the sliding window (in the output table) does not exceed the length of the query sequence.
- 80-mer search: now shows appropriate maximum query windows (length of sequence - 79).
- 8-mer search: now shows appropriate maximum query windows (length of sequence - 7).
- Update (April 17, 2020): Remove E-Value field for 8-mer search page (E-values not appliable for 8-mer searches).
- Update (April 17, 2020): Changed the "default" E-Value from "1" to "10", in the full-length and 80-mer search options.
- The reason for the change is to reflect the standard E-value referenced in the FASTA A software (Pearson, 2000) and also commonly used in BLAST searches (where "10" is the default E-value). However, multiple E-value options remain available in the COMPASS tool (100; 1; and under); users have the possibility to select the E-value that best suits their specific needs.
- Issue: limitations in the maximum length of input sequences (error message: "Your input sequence is too large")
- Solution (June 29, 2020): input sequence length cut off set at 3,500 aa (more than twice the current largest sequence in COMPARE). Large sequences now accepted and running properly in the 3 search options (note that 80mer and 8mer searches may take longer due to higher number of sliding windows to be processed).
- New Feature (July 29, 2020): COMPASS now offers a visualization option to view results in a color-coded graphic display.
- Issue: When the returned alignment for a library sequence is too large (more than 3,000 characters), the tool outputs "No alignment found.".
- Solution (August 7, 2020): When the returned library sequence alignment is too large, the returned alignment for each result is limited to three rows due to large amount of characters. User is recommended to refer to the raw output to see the full alignment result.
- Remove Significance Limits table in Full Length and 8-mer searches (August 7, 2020): Significance Limits table under "Recap and table details" tab is only present on 80-mer search.