Literature and Los Angeles

What's Changed

• Improve performance by removing regex from indel location analysis by @Colelyman #182
• Fastq output produced by --fastq_output now includes the inserted bases. Previously, a string like "DEL= INS=78(1) SUB= " would indicate a 1bp insertion at site 78. This update outputs strings like "DEL= INS=78(1+G) SUB= " with the insertion described as a plus character followed by the inserted bases. (2f84dd0)
• Adds features for providing aligned bams as input to CRISPRessoPooled and for a faster demultiplexing when amplicons and genome are provided. The added parameters are:
  • --aligned_pooled_bam: Path to aligned input for CRISPRessoPooled processing. If this parameter is specified, the alignments in the given bam will be used to demultiplex reads. If this parameter is not set (default), input reads provided by --fastq_r1 (and optionally --fastq_r2) will be aligned to the reference genome using bowtie2. If the input bam is given, the corresponding reference fasta must also be given to extract reference genomic sequences via the parameter --bowtie2_index. Note that the aligned reads are paired-end seqenced, they should already be merged into 1 read (e.g. via Flash) before alignment.
  • --demultiplex_only_at_amplicons: If set, and an amplicon file (--amplicons_file) and reference sequence (--bowtie2_index) are provided, reads overlapping alignment positions of amplicons will be demultiplexed and assigned to that amplicon. If this flag is not set, the entire genome will be demultiplexed and reads with the same start and stop coordinates as an amplicon will be assigned to that amplicon.

Minor bug fixes:

• Fix deprecated pandas indexers (eea442a, f4b6cfc)
• Update ylabel_values -> y_label_values by @swrosati in #174
• Allow mixed-case prime-editing input (e999079)
• Fix int bug for CRISPRessoPooled n_reads (ef15cae)

Full Changelog: v2.2.6...v2.2.7

Basepairing Bern

Various improvements and bug fixes:

• Add param --plot_center to allow custom plots centered at a given point in plotCustomAllelePlot script (ecf23ef)
• All sub-CRISPResso runs are run with 1 process in Batch, WGS, Pooled, etc. Because we added multiprocessing capabilities to CRISPResso (the plotting part) we thought it would be slick for CRISPRessoPooled to run sqrt(n_processes) CRISPResso processes with sqrt(n_processes) processes each. Unfortunately, sqrt(n) isn't a really useful number for the number of processes people usually run (e.g. 2 or 3 or even 8), and a lot of the CRISPResso processing isn't enabled to take advantage of multiprocessing (e.g. the alignment step), so processes were being wasted. So we reverted back to having n_processes CRISPResso processes, each with 1 process. (a923a7c)
• Remove version checks for numpy and seaborn (90b43ea)
• Fix allele plotting error for plot 5 referring to uninitialized y_max variable (53197e6)
• Fix unicode errors for bam read/write (8196b6a)
• Convert columns in nucleotide count and modification tables to numeric for PE analysis (cabebbe)
• Make loggers module-specific so matplotlib debug doesn't get spewn out in the CRISPResso log (c2bdd96)
• Add unit tests (3e6c281)

Full Changelog: v2.2.5...v2.2.6

Immunity from Bonneville

Fixes bug when sequencing reads are much longer than the given reference sequence.

Mutant Maricopa

This release adds an additional parameter --assign_ambiguous_alignements_to_first_allele. For ambiguous alignments, setting this flag will force them to be assigned to the first (as provided by the references -a first and then -e second) amplicon. Thus, no reads will be discarded as 'ambiguous' and all reads will be counted once in the analysis.

In addition, batch summaries are produced for amplicons present in only one sample.

Collateral Cardston

Fixes database_id bug

Large Honolulu

For some reason some of the previous commits resolving problems with filterFastqs weren't picked up in v2.2.1. So I'm hoping they'll be included here.

Sequence Length Salt Lake

• More unicode bug fixes for filtering fastqs
• CRISPRessBatch now outputs summary of splicing/frameshift mutation status

Matches Sanpete

Python 3 release

• Incorporates updates and changes from python2 up to this point
• Adds multiprocessing to CRISPRessoBase to parallelize image generation to speed up results
  • CRISPRessoBatch, CRISPRessoPooled, and CRISPRessoWGS allocate processes to sub-CRISPResso commands so that sqrt(n_processes) sub-commands are run, each with sqrt(n_processes) unless plotting is turned off (via --suppress_report or --suppress_plots in which case n_processes are run, each with 1 process.
• The crispresso_info dictionary containing run information is saved as json so it can be read across versions of python and by other programs (e.g. R). The dict structure of the object has also been changed to be more navigable and hierarchical.

Because of changes to crispresso_info, this version of CRISPResso will not be able to finish incomplete runs (e.g. checkpointing of CRISPRessoPooled) that were started by previous python2 versions of CRISPResso, or aggregate or access information from previous runs (e.g. using CRISPRessoAggregate, CRISPRessoCompare, or the custom python plotting scripts in scripts). However, even if we were to have stuck with pickle format for crispresso_info, the python2 and python3 versions were incompatible anyway. So we figured this was a good time to move toward a better format.

Lentiviral Cache

Fixes bug for CRISPRessoPooled analyzes of many amplicons where samtools sort writes status updates that can't be parsed as reads.

Single Guide Washington

This release includes:

• Updates to CRISPRessoPooledWGSCompare, used for comparing multiple amplicons in CRISPRessoWGS or CRISPRessoPooled experiments 48d6c87
  • CRISPRessoPooledWGSCompare now produces html report linking to sub-CRISPRessoCompare reports
• Addition of a script for custom allele plotting 947fbab