Flexible cell assignments simpler

[ekatsevi]

• I added the optional logical argument remove_cells_w_zero_or_twoplus_grnas
  to run_qc(), which is relevant for low-MOI analyses with control_group = "nt_cells".
  If set to TRUE, then cells with zero or two or more gRNAs are removed from the analysis, as before.
  If set to FALSE, then cells with multiple gRNAs are kept in the analysis if
  they contain at most one targeting gRNA. For these cells, a perturbation identity
  can be unambiguously defined, and their inclusion increases the number of cells
  available for downstream analysis.
• The data frames outputted by the calibration check, power check, and discovery
  analysis have two additional columns called n_trt and n_cntrl, giving the
  number of cells used in the treatment and control groups, respectively.
• I have resolved the plotting bug described in this issue.
• I have added a check in construct_cis_pairs() that grna_target_data_frame
  has columns chr, start, and end, resolving this issue.