v1.2.2 - 3/15/2024

The correct files to download are either falco-1.2.2.zip or falco-1.2.2.tar.gz

Please do not download the "Source code" files automatically generated by GitHub, as they do not contain all files necessary for compilation.

What's Changed

• Fix error in parsing gzip file by @y9c in #48
• Added direct support for reading BAM files to improve efficiency in some situations by @masarunakajima in #53

New Contributors

• @y9c made their first contribution in #48
• @masarunakajima made their first contribution in #53

Full Changelog: v1.2.1...v1.2.2

v1.2.1 - 9/15/2022

The correct files to download are either falco-1.2.1.zip or falco-1.2.1.tar.gz

Please do not download the "Source code" files automatically generated by github, as they do not contain all files necessary for compilation

This is a quality-of-life release that fixes two serious bugs in 1.2.0

bug 1 (from issue #40) : falco stalls if the default configuration files are not used (e.g. when using conda)
bug 2: the file format type was not being recognized properly when using multiple inputs of different file types.

v1.2.0 - 9/11/2022

The correct files to download are either falco-1.2.0.zip or falco-2.1.0.tar.gz

Please do not download the "Source code" files automatically generated by github, as they do not contain all files necessary for compilation

New in v1.2.0:

• Custom file names for summary, data and HTML reports can be provided by users
• More detailed documentation of parameters. Much of it was directly copied from FastQC v0.11.9, with flags that are not yet implemented being specifically stated as such

Bugfixes:
#39 : We fixed a bug where reads of length >1M were crashing falco. The program now allows reads of about 64M bases, but note that, since falco collects data for every read position, memory will scale accordingly. Plan the computing resources ahead!

Other changes:

• The Conda release now contains the default contaminants, adapters and limit files under the opt/falco-1.2.0/Configuration directory when installing through it, consistently with FastQC.

v1.1.0 - 09/02/2022

The correct files to download are either falco-1.1.0.zip or falco-1.1.0.tar.gz

Please do not download the "Source code" files automatically generated by github, as they do not contain all files necessary for compilation

New in v1.1.0:

• Subsampling reads to improve speed for large files

Bugfixes:
#37 : adapter plot now shows range corresponding to max read length
#36 : support for multiple fastq files in a directory
#23 : html report for k-mers had a bug that is fixed

Other changes:

• modified code so falco can build with older compilers still common on some systems
• fixing a bug where the output directory was being duplicated if file is within a directory
• fixing a bug where quiet falco still displayed full progress bar
• less frequent updates to progress bar make the code faster

The added functionality is the "subsample" option, which uses every k-th read. For large files, this will give the same results (converging at the usual rate), but is faster. Thanks to @y9c for the suggestion (Issue #35).

v1.0.0 - 05/24/2022

The correct files to download are either falco-1.0.0.zip or falco-1.0.0.tar.gz
Please do not download the "Source code" files automatically generated by github, as they do not contain all files necessary for compilation

New features in 1.0.0:

• Progress bar for FASTQ, gzipped FASTQ and SAM inputs.
• Falco now runs in less than 100 MB for SAM input

Bugfixes
#18 : falco now automatically identifies quality encoding with an identical approach to FastQC
#20 : falco now recognizes filenames independent of lower/uppercase
#23 : falco can now handle adapters of varying sizes
#25 : consistency with FastQC by rounding per base sequence quality distribution to double
#24 : falco can now handle different types of newline
#26 : falco no longer crashes if a read is empty or too short
#6 and #33 : Major fixes in BAM and SAM input processing

Major thanks to @Shelestova-Anastasia for bringing up the problem and contributing to the fixes for #23 , #24 , #25 and #26 !!

v0.3.0 - 9/9/2021

The correct files to download are either falco-0.3.0.zip or falco-0.3.0.tar.gz
Please do not download the "Source code" files automatically generated by github, as they do not contain all files necessary for compilation

This release adds new functionality to falco and addresses several bugs reported in the issues section, listed below

#14 : When using the --nano flag, the basic statistics file will display "Oxford Nanopore" instead of Illumina. Analysis is still done for Illumina sequences. We need a fast5 parser in C++ (or write one from scratch) to allow fast5 support, which is not yet available

#15 : The k-mer content module was implemented, but is still missing p-value calculations. For now it approximates FastQC behavior, and in the next release it will be identical. We also show 10 k-mers instead of 6 in FastQC since we can erase some of them through plotly to not make the plot so dense

#16 : Fixed a bug where, if the adapters file was not provided (which is the case in the Conda release), adapters were not being counted.

#17 : Fixed several bugs in the overrepresented sequence "hits" comparison with the contaminant list, including allowing partial overlap between sequence and contaminant.

v0.2.4 - 1/8/2021

The correct files to download are either falco-0.2.4.zip or falco-0.2.4.tar.gz
Please do not download the "Source code" files automatically generated by github, as they do not contain all files necessary for compilation

This release resolves the following issues:

• Memory usage for uncompressed FASTQ files has been drastically reduced by replacing memory mapping with standard C IO. versions <= 0.2.3 use most of the available memory in the system, whereas this version only uses about 92 MB for short reads and under 400 MB for long reads
• Base groupping has been implemented for per base sequence content and per sequence N content modules, which also drastically reduced the output size of the HTML file and the text report, especially in long-read files. We opted to leave sequence length distribution without base groupping.
• Fixed a bug in adapter content where certian long-read datasets were outputting nonsensical values.

v0.2.3 - 2020/12/22

The correct files to download are either falco-0.2.3.zip or falco-0.2.3.tar.gz
Please do not download the "Source code" files automatically generated by github, as they do not contain all files necessary for compilation

This release resolves the following issues:

• HTML reports were fixed in cases where certain positions in the tile x position matrix did not contain any values. Their z-scores are defaulted to zero
• When multiple files are given as input, previous versions were causing problems where only the last given dataset had output files generated. This release resolves this issue
• Color heatmaps in HTML the "per tile sequence quality" are now more consistent with FastQC, where red and blue shading only depend on the absolute z-score value, and not on the z-score value distribution within the dataset. This makes the meaning of "red" or "blue" clearer, and consistent with whether the file passes or fails the module assessment.

v0.2.2 -2020/09/13

The correct files to download are either v0.2.2.zip or v0.2.2.tar.gz
Please do not download the "Source code" files automatically generated by github!

This release addresses some issues found on verision 0.2.1 that were reported by users:

Issues #4 and #7 : The format of fastqc_data.txt was corrected to be parsable by MultiQC version 1.9

Issue #8 : Falco now displays a hot-cold heatmap with absolute color values that do not depend on the z-score distribution of the samples: tile-position qualities will be shown as light blue when quality is around 0, dark blue when quality approaches +10, and red when it approaches -10. This allows the heatmap to be usable in highlighting the specific tiles that failed the QC test (i.e. the ones closest to the red color)

Issue #9 : A bug was fixed where falco would crash if a larger read was shown later in a dataset with the same tile value as a previous shorter read.

This release will soon be added to bioconda.

v0.2.1 - 2019/12/12

This release gets rid of the template.html file in the Configuration directory. An html template is no longer required for the program to run. Major changes in the source code also make it more readable. All "magic" constants used throughout the program are centralized in the Constants namespace in the FalcoConfig.hpp file.