v0.3.0

• mutations are mostly unfiltered
• exon and gene level counts are separated
• HTseq fixed to calculated correct strandedness

v0.2.0

Strategy change

1. Reads aligned with HiSAT2
2. Alignments split into:
   1. reads aligning to RNA fragments originating from + strand
   2. reads aligning to RNA fragments originating from - strand
3. collect non-duplicate readids using fastuniq (Can used picard MarkDuplicates to remove duplicate reads from the BAM as well... but sticking to fastuniq to be inline with the TimeLapseSeq paper)
4. alignments to fragments originating from + strand and - strand are filtered for only non-duplicate readids
5. non-duplicate alignments to fragments originating from + strand are used to call T-to-C mutations
6. non-duplicate alignments to fragments originating from + strand are used to call A-to-G mutations
7. using T-to-C mutations from step E. split BAM from step 2.i. into mutated and unmutated BAMs
8. using T-to-C mutations from step E. split BAM from step 2.ii. into mutated and unmutated BAMs
9. merge mutated BAMs from steps 7. and 8. to produce mutated BAM per sample
10. merge unmutated BAMs from steps 7. and 8. to produce unmutated BAM per sample

Other changes

• Major changes to the wrapper script
• changes to output folder initialization
• master snakemake script is also a submitted job on the cluster ... no need for screen

To Do

• Gene level counting with HTSeq
• Differential expression with DESeq2(possibly insert counts from HTSeq into CCBR_Pipeliner)

v0.1.0

• First working version.