Add support for cell/sample barcodes and demultiplexing logic to calib-dedup step

[nsyzrantsev]

There is calib-dedup step. You should to do:

• Implement a pyumi step to run umi preprocessing alone #13
• Add supporting for UMI (molecular barcode) pattern. Move UMI to the read start and use it in deduplication (already done);
• Add cell barcode support in pyumi step #37
• Add supporting for SMPL (sample barcode) pattern. Use SMPL to split (demultiplex) into files and check it in the whitelist; Add new parameter --demultiplex. If you specify it fastq will split into files by SMPL barcode

[mikessh]

To add: glue multiple UMI groups into a single sequence in case of adapters with complex UMI structure like NNNTNNNTNNNTCTT

[mikessh]

Add group_type <> group correspondence dictionary as a parameter. By default UMIi where i is an integer, CBi and SBi should be considered. E.g. the default for this parameter can be

bc_type_r1 = {"UMI1": "UMI", "CB1": "CB", "SB1": "SB"}

[nsyzrantsev]

@yuliajk, I think is a good idea to move all code with umi preprocessing into a new pyumi step (calib-dedup step will contain code with running only calib tool). For more information see: #13

[yuliajk]

Agree