Version 0.2.1

Updates from v0.1.10:

• Influenza B virus is now supported. This necessitated several changes, hence the rationale for incrementing the minor version to 0.2.x. Since the bioinformatic strategy remains the same, this was not deemed to be a major update to version 1.x.x.

• Implementing influenza B support required changes to the formatting of headers in the FluViewer db. Sequences must now be annotated with species in addition to segment and subtype. A new default FluViewer db has been provided that includes some recent influenza B virus reference sequences.

• Implementing influenza B support also changed how consensus sequences are quality-filtered based on their length. Previously, ranges of permitted sequence lengths were hard-coded for each segment. Since influenza A and B viruses have evolved different length distributions for their segments, those ranges led to the rejection of some valid influenza B segment consensus sequences. Now, the range of accepted segment lengths is determined from the FluViewer db provided by the user. Before each analysis, FluViewer scans the contents of the provided to ensure the headers have proper formatting and to gather the length distribution for each species/segment/subtype combination in the db. Consensus sequences generated by FluViewer must not be more than the length tolerance shorter than the shortest example of that species/segment/subtype in the db. Similarly, they must not be more than the length tolerance longer than the longest example of that species/segment/subtype in the db. The length tolerance is a percentage that can be set by the user using the -t option (default is 1%).

• Implementing influenza B support also changed how mixed infections are handled. FluViewer will fail any specimen if more than one influenza species is detected, more than one influenza A HA or NA subtype is detected, or more than one influenza B lineage is detected. There is a new, experimental “allow mixtures” mode that can be activated with the -m flag. If activated and a mixture of species or subtypes/lineages is detected, FluViewer will proceed with its analysis and generate only consensus sequences for species/segment/subtype combinations detected with a subtype other than “none”. For example, in a mixture of influenza A H1N1 and influenza A H3N2, consensus sequences would be generated for H1, H3, N1, and N2 segments, but not for any of the internal segments (ie PB2, PB1, PA, NP, M, and NS). Intrepid users may attempt to create a custom db where internal segments from sufficient divergent lineages have been given their own “subtype” codes to attempt recovery of more segments from mixed infections, but proceeding with caution is advised and this is not currently supported.

• Normalization and downsampling of reads by BBNorm is no longer part of FluViewer. This was included in previous versions to address uneven amplification by commonly used amplicon-based lab protocols. This step was resource intensive, but might be unnecessary for many users. As such, it was removed, and downsampling/normalization is now left at the discretion of the user, much like trimming of amplification primers and sequencing adapters.

• Runtime inputs provided by the user are now parsed for characters reserved by the system (eg |:;/). This is to prevent injection of destructive/malicious code in runtime parameters.

Version 0.1.10

Updates from v0.1.9:

• Fixed a bug in the command line interface where not specifying a number of threads raised an error
• When generating the mapping reference, the reference sequence genotype was only transferred for spans of Ns between contigs, leaving Ns in the mapping reference if those Ns existed within the scaffold sequence. Now, Ns inside contigs are also replaced by the corresponding positions of the mapping reference.
• The amount of memory allocated for BBNorm's kmer counting hash can now be set through the FluViewer command line interface using the -M argument
• Fixed an edge case bug where coincidental reversed alignments between the scaffold and best-matched reference sequence introduced unexpected negative values into string slice indices, resulting in the full best-matched reference being concatenated onto the mapping reference. Now, reversed alignments between scaffold and best-matched reference sequence are discarded. The length of each consensus sequence is also checked to see if it falls within the expected range for its genome segment to catch any other edge cases like this.
• The length of each sequence in the reference database is checked to make sure it falls within the expected range for its genome segment.
• Fixed a bug where variant positions were not indicated on the depth of coverage plots by blue lines.

Version 0.1.9

Changes since v0.1.8:

• Installation through YAML because of dependency conflicts
• Fixed bug with masking of ambiguous positions

Version 0.1.8

Major update to FluViewer:

• No more align or assemble modes... the best aspects of both are combined by implementing a contig scaffolding algorithm
• Variant calling using FreeBayes due to artefacts and errors with bcftools
• New report format

Version 0.0.2

Small fixes:

• Version 0.0.1 created a TSV file called "blah.tsv". This was a carry-over from trouble-shooting that I forgot to remove! It is no longer created.
• In cases where flu contigs were created, but all of them had insufficient positions with ATGC and minimum depth of cov ("sequenced bases") to achieve minimum segment coverage, an empty consensus_seqs.fa file was created. This led to an empty input file error message when it was opened later to count sequenced bases for the report. There is now a check before writing the consensus sequences after low coverage positions are masked to make sure the file would not be empty.

Version 0.0.1

First stable release.

Version 0.0.0 alpha

Base FluViewer code. No align mode.