Standardizing inputs to AIRR format

[dltamayo]

With multiple potential sources of TCR count tables, the pipeline should be able to handle inputs beyond Adaptive. Particularly relevant is taking outputs from CellRanger for scRNA TCR sequencing, to enable pseudobulking of single-cell data (see #33).

Bare minimum columns:

• full nucleotide sequence
• full nucleotide aa sequence (can infer from full nuc seq)
• CDR3 nucleotide (can infer from full nuc seq)
• CDR3 aa sequence (can infer from full nuc seq)
• count
• VJ identifications

We will need to agree on a standard to conform Adaptive and any other imported count tables to; AIRR is a possible standard widely-adopted by the open source community. Other pipelines have been certified as AIRR-compliant, and I will investigate how we can conform our inputs to this standard.