Samples for analysis of CD99 expression on B cells subsets were stained and analyzed using validated laboratory procedures. Antibody combinations are listed in supplemental table 2 . Samples were stained with the antibody cocktail for 15 minutes, lyse-fixed (10% ammonium chloride lyse + 0.25% formaldehyde) for 15 minutes, and washed with PBA solution (phosphate buffered saline + 0.3% bovine serum albumin + 0.1% percent sodium azide preparation) prior to acquisition. The samples were acquired on standardized FACS Canto 10-color instruments (BD Biosciences, San Jose, CA), and analyzed using custom Woodlist software (gift of Brent Wood, University of Washington).
For plasma cell analysis, all samples were stained using a standardized procedure as described previously for Memorial Sloan Kettering single-tube 10-color method, and acquired on FACS Canto 10-color instruments30 , 32 (link). The abnormal plasma cell population immunophenotype was first identified by Memorial Sloan Kettering single 10-color tube assay using antibody combination described in Table 2. The CD99 was then substituted for kappa FITC/lambda PE light chain antibodies in the investigational panel for evaluation of CD99 (supplemental table 2 ). The abnormal plasma cells were gated based on known aberrant antigen expression from the current 10-color tube. CD99 expression was not used in gating to avoid bias.
For plasma cell analysis, all samples were stained using a standardized procedure as described previously for Memorial Sloan Kettering single-tube 10-color method, and acquired on FACS Canto 10-color instruments30 , 32 (link). The abnormal plasma cell population immunophenotype was first identified by Memorial Sloan Kettering single 10-color tube assay using antibody combination described in Table 2. The CD99 was then substituted for kappa FITC/lambda PE light chain antibodies in the investigational panel for evaluation of CD99 (