Autologous Antibody Capture to Enrich Immunogenic Viruses for Viral Discovery

Oude Munnink, Bas B.; Jazaeri Farsani, Seyed Mohammad; Deijs, Martin; Jonkers, Jiri; Verhoeven, Joost T. P.; Ieven, Margareta; Goossens, Herman; de Jong, Menno D.; Berkhout, Ben; Loens, Katherine; Kellam, Paul; Bakker, Margreet; Canuti, Marta; Cotten, Matthew; van der Hoek, Lia

To identify sequences which are enriched by antibody capture, input and enriched sequences were compared using a Python algorithm xcompare (source available on request). The algorithm is comprised of the following steps: the identification of identical or nearly-identical sequences within the input subset was performed by creating a custom BLAST database [19] comprising all the reads derived from the input sample and subsequently performing a BLAST search within this database towards its own sequences. Via this method a list of closely related sequences could be created which were extracted and aligned via MUSCLE 3.8.31 [20, 21] (maximum number of iterations: 1; diagonal optimization enabled) and both consensus sequences and unique sequences were joined into a unique fragment library. In turn, this unique fragment library was converted into a second custom BLAST database, against which the reads obtained from the enriched sample were compared. For the BLAST comparison the following settings were used: E-value threshold 3E-60 for within the input sample library and 3E-25 for between the enriched and input samples library; word size: 11; match/mismatch scores 1/-2; gap existence/ extension penalty: 5/2.