Vidjil add-on for MRD quantification of samples processed using the EuroClonality-NGS protocol

Abstract

Assessment of minimal residual disease in acute lymphoblastic leukemia by immune repertoire NGS requires spiking CDR3 sequences at known quantities into the patient's sample. Recently, the EuroClonality-NGS group released one of the most comprehensive protocols for this purpose. ARResT/Interrogate is a closed-source software for processing these NGS libraries, developed by this same group. Vidjil, an open-source alternative, currently cannot handle libraries prepared using this protocol. Here, we present a Vidjil add-on to solve this issue. EuroClonality-NGS prepared samples analyzed with Vidjil and ARResT/Interrogate were highly concordant (r = 0.998) and presented low error (root-mean-square error, RMSE = 0.112).

Keywords: T‐cell receptor; Vidjil; immune repertoire; immunoglobulin genes; leukemia; minimal residual disease.

FIGURE 1

EuroClonality versus ARResT/Interrogate: IG/TR marker quantification and cIT‐QC identification. Analysis of bone marrow NGS files from the original cIT‐QC publication [6]. Samples S1‐S32 correspond to diagnostic B/T‐ALL, while S33‐S48 to aplastic posttreatment. (A and B) Linear regression of the log₁₀ cell frequency for IG/TR clonotypes calculated by ARResT/Interrogate and Vidjil. Pearson's correlation coefficient represented by r, and root‐mean‐square error by RMSE. Vidjil analysis without (A) and with manual clonotype clustering (B). (C, D, and E) ARResT/Interrogate and Vidjil comparison per sample of total cIT‐QC read count (C), total IG/TR rearrangement read count (D), and total cIT‐QC read frequency (total cIT‐QC reads/total V(D)J reads) (E). Green and black arrows indicate samples where a cIT‐QC sequence was missing in both software tools and only in Vidjil, respectively.