In January 2022, I started a permanent research position at INRIA, the French National Research Institute for Computer Science and Automation; see my new website on http://karel-brinda.github.io.

Prior to that, I was a postdoctoral researcher in the Laboratory of Michael Baym in the Department of Biomedical Informatics at Harvard Medical School, and in the Center for Communicable Disease Dynamics in the Department of Epidemiology at the Harvard TH Chan School of Public Health. I received my Ph.D. in computer science at Université Paris-Est, while researching also at Institut Curie and Institut Cochin, supervised by Gregory Kucherov and Valentina Boeva. Prior to that, I received a bachelor and masters degrees in mathematical computer science at the Faculty of Nuclear Sciences and Physical Engineering at the Czech Technical University in Prague, under the supervision of Edita Pelantová.

The fundamental goal of my research is achieving rapid, robust and affordable diagnostics and real-time surveillance of infectious diseases at points of care. To do so, I develop novel algorithms, data representations, software and databases for computational genomics, which are then provided to the scientific community as building blocks for larger efforts. I am particularly interested in the problem of search across all existing sequence data, as well as unlocking the full potential of portable genomic technologies, such as nanopore sequencing and CRISPR-Dx.

Genomic Neighbor TypingMy recent work includes Genomic Neighbor Typing, a method that can identify antibiotic resistance within 10 minutes of nanopore sequencing (and 4 hours of sample collection from metagenomic sputum samples); by similarity genome search across data from epidemiological surveillance, it can rapidly infer the likely resistance profile of a sample from its closest known relatives. Genomic Neighbor Typing built on my contributions to computational genomics, including simplitigs as an efficient and scalable representation of de Bruijn graphs (and ProphAsm for their computation), the ProPhyle metagenomic classifier, and ProphEx for k-mer indexing using the Burrows-Wheeler Transform.

My other work includes spaced seeds for metagenomic classification, online consensus and variant calling for a rapid analysis of streamed NGS data (Ococo), a format and a toolkit for evaluating read mappers (RNFtools), and the concept of dynamic mapping for better read mapping and variant calling. I also occasionally contribute to large projects such as Bioconda or Snakemake and develop methods for automatic generation of tactile maps for blind users (Blind Friendly Maps).

Selected publications