Merging new v1.1 features and fixes

[skchronicles]

Changelog

• Fix: increasing allocated memory, lscratch, and walltime requirements for several rules
• Feature: adding Dockerfile for new custom multi-threaded version iCount
  • cutadapt and demultiplex commands now utilize multiple threads
  • Updating adapter trimming rules and demultiplex rules to use 16 threads
• Fix: removing problematic usage of set +e from any shell commands
• Chore: reformatting all shell commands for increased readability
• Fix: adding fix for samtools view -H standard input pipe failures from premature SIGPIPE
• Chore: moving fastQValidator to a common shared area
• Fix: adding fix to ensure .snakemake directory/lock is creating in the pipeline's output directory
• Fix: moving away from depreciated built-in dynamic() to avoid DAG re-evaluation when the pipeline runs
  • bug was causing the pipeline to re-run extremely compute intensive steps like cutadapt, demultiplex, and novoalign all over again in un-related downstream rules (even if they had already run successfully)
  • should results in an immediate and significant decrease in overall run time
• Feature: adding new configurable option, splitSampleNChunks, to spilt up sample into a deterministic number of chunks
  • Can be used to speed up compute heavy tasks like alignment as replace to the previous dynamic() method
  • Adding safe guard to prevent the user from setting this to an insanely high number
• Fix: fixing input to rule qc_screen_validator which was causing fastqValidator to fail, also adding a new option to fastqValidator to reduce memory profile of tool
• Fix: adding new method to extract the correct UMI delimiter to prevent error in umi_tools dedup
• Fix: merging trimmed and untrimmed reads from iCount cutadapt's output, previously any 3' adapter trimmed reads were not being used in any downstream steps
• Fix: updating run_snakemake.sh so it can be invoked from anywhere on the filesystem
• Fix: setting default DEMETHOD in config to none