1Genotypic resistance interpretation systems for the prediction and interpretation of HIV-1 2 antiretroviral resistance are an important part of the clinical management of HIV-1 infection.3 Current interpretation systems are generally hosted on remote webservers that enable clinical 4 laboratories to generate resistance predictions easily and quickly from patient HIV-1 sequences 5 encoding the primary targets of modern antiretroviral therapy. However they also potentially 6 compromise a health provider's ethical, professional, and legal obligations to data security, 7 patient information confidentiality, and data provenance. Furthermore, reliance on web-based 8 algorithms makes the clinical management of HIV-1 dependent on a network connection. Here, 9 we describe the development and validation of sierra-local, an open-source implementation of 10 the Stanford HIVdb genotypic resistance interpretation system for local execution, which aims 11 to resolve the ethical, legal, and infrastructure issues associated with remote computing. This 12 package reproduces the HIV-1 resistance scoring by the web-based Stanford HIVdb algorithm 13 with a high degree of concordance (99.997%) and a higher level of performance than current 14 methods of accessing HIVdb programmatically.
15
INTRODUCTIONGenotype-based prediction of human immunodeficiency virus type 1 (HIV-1) drug resistance is an 16 important component for the routine clinical management of HIV-1 infection [1, 2]. Detecting the 17 presence of viruses carrying mutations that confer drug resistance enables physicians to select an 18 optimal drug combination for that patient's treatment regimen. Furthermore, genotyping by bulk # Corresponding author, apoon42@uwo.ca sequencing is a cost-effective alternative to the direct measurement of drug resistance from cultur-20 ing virus isolates in a laboratory [3]. Provided access to affordable bulk sequencing at an accredited 21 laboratory for clinical microbiology, the interpretation of HIV-1 sequence variation is the primary 22 obstacle to utilizing resistance genotyping for HIV-1 care. Fortunately, there are several HIV-1 23 drug resistance interpretation algorithms that can be accessed at no cost through web applications 24 or services hosted by remote network servers, such as the Standard University HIV Drug Resis-25 tance Database (HIVdb) [4], Agence Nationale de Recherche sur le SIDA (ANRS) AC11 [5], and 26 Rega Institute [6] algorithms. The Stanford HIVdb interpretation system can be accessed either 27 through a web browser at http://hivdb.stanford.edu/hivdb or programmatically through its Sierra 28 Web Service [7], which requires the transmission of an HIV-1 sequence from a local computer 29 over the network to the remote server. This is a convenient arrangement for clinical laboratories 30 because there is no need to install any specialized software, web browsers are ubiquitous and most 31 users are familiar with submitting web forms. 32 On the other hand, there are a number of disadvantages to accessing interpretatio...