Major updates - software versions, new wrappers, documentation

This release fixed issues brought up by users and at the same time updated the software versions to be more current.

Issues resolved

• #49 was addressed by adding to the documentation to clarify the SNP filtering step
• A comment was added to the Scrublet docs to account for some users needing to set the NUMBA_CACHE_DIR for #45
• annoy was downgraded to v1.15 to deal with some issues that were observed on some clusters with annoy 1.17 #44
• Wrapper scripts were introduced for checking the chromosome encoding in the bam, vcf and fasta files (where appropriate) to ensure they are the same; docs were updated to call these wrappers instead of the commands directly #39
• Docs were updated to fix the running code provided for Demuxalot from singularity exec Demuxafy.sif python Demuxalot.py ... to the correct call: singularity exec Demuxafy.sif Demuxalot.py ... #28
• Added a note about the usefulness of not including --SAMPLE_FILE in the AssignCellsToSamples to identify potential sample swaps #27

Software updates:

• R was updated to v4.4.0 (from 4.2.2)
• python to v3.9.7 (from 3.7.2)
• Seurat updated to v5.1.0
• Dropulation updated to v3.0.1 (from 2.5.4) which has more tools for implementing and demultiplexing
• Souporcell v2.5 (from v2) -> now supports hisat2 mapping and allows for implementation on newer python versions
• Demuxalot v0.4.1 (from 0.2.0) with help from @mschilli87 #46
• Note that solo was wrapped into scvi but I have not tested that version or identified how to alter the scripts to adapt to this change so the solo included here is the last solo version compatible with python 3.9 provided outside of scvi
• Update the scds R script to account for situations where there are less than 500 variable genes (by default it causes an error if not accounted for in bcds call)

Thanks to @mschilli87 for the help in identifying issues and helping resolve them.

v2.1.0 zenodo

Zenodo requires a new release after linking for it to work

v2.1.0

Major update to fully include demuxalot and dropulation and add additional functionality. Updates include:

• A software selection tool is now included so that users can provide the dataset characteristics and it will indicate the recommended software combination and the expected number of doublets and singlets
• Demuxalot can now be provided to the Combine_Results.R script
• Dropulation can now be provided to the Combine_Results.R script
• The Combine_Results.R script has been updated to now provide summaries of the number of singlets and doublets as well as the number of droplets annotated to each donor
• Documentation now includes a Tutorial that users can follow
• I have added a link to a location to get a fasta file for the Test dataset for running souporcell (issue #22)
• I have added biopython to the image so that fasta files can be gzipped (issue #23)
• The time to run each step of each software has been added to the documents
• The Combine_Results.R script now can handle single softwares and provides the results in a consistent manner as well as bar plots of the droplet annotations.

Some other small updates to scripts and documents were included as well. Also fixed some typos in docs.

v2.0.2

Only update is scanpy is imported in DoubletDetection and scrublet scripts - a previous oversight required for importing h5 files.
Singularity image was updated and the links for downloading the updated image and md5 file as well.

v2.0.1

This version has no updates to any Demuxafy scripts but does update the vartrix version from 1.1.13 to 1.1.22 to resolve some bugs.

v2.0.0

Updates:

• Updated Combine_Results.R script to provide unassigned if multiple demultiplexing software call a singlet but do not match or return unassigned
• Updated upset plots in Combine_Results.R script to provide more information and functionality
• Fixed order of software in resulting upset plot for easy comparison across pools
• Provide more messages in Combine_Results.R scripts to notify when unexpected behavior of cell assignment is observed
• Add functionality to provide a filtered barcode list to scds, scDblFinder, DoubletDetection and Scrublet if don't want to process all the barcodes in the filtered matrix or h5 file
• Updated issue with h5 handling in scds and scDblFinder scripts
• Fixed missing -N $N in Vireo command when only some genotyped individuals
• Provided filtered vcf files of common SNPs from 1000G for users
• Provided documentation about how to integrate Demuxafy results into single cell objects in R (Seurat) and python (AnnData)
• Add singularity binding reminders below first singularity commands on each page

v1.0.3

Fix a couple typos/bugs in the combining script
Update versions to use for readthedocs
update souporcell summary script to count unassigned separate from doublets

v1.0.2

This release fixes a bug that happened when combining results from different methods that resulted in some empty entries in the final assignment. It also provides more useability by expanding input to also be able to use 10x style h5 files and provides new visual for the combined results to enable easy visualization of droplet assignments for combinations of methods.

v0.0.4

Finalized the Combine_Results.R script:

• Includes intersectional methods for calling combined cell type from multiple softwares
• Includes method for providing common ids when only non-reference SNP genotype methods are used

v0.0.3

This is the first viable release that has functional documents and a functional singularity image. There are still some holding places in the documentws (ie citation and recommended software combination)