Research

STORM Research Group

Our primary research goal is to unlock transformative applications in medicine and health by substantially improving the scalability and accessibility of biological data analysis (e.g., omics data). To achieve this goal, we design algorithms and co-design hardware and software for more accurate, high-performing, and energy-efficient omics analysis.

Our STORM Research Group at the University of Maryland spans algorithm design, hardware-software co-design, and AI-driven methods for biological data analysis. For a complete list of publications, please see our publications page, Google Scholar, DBLP, or ORCID.

Research Directions

Multimodal Systems for Biological Data Analysis

Modern biological datasets are large, noisy, and heterogeneous. To identify meaningful insights from complex biological data, combining complementary data modalities can help resolve certain complexities. To this end, we aim to develop multimodal systems that use different types of biological data (e.g., raw electrical signals, basecalled sequences, spatial and image data).

Hardware-Algorithm-Software Co-Design for Portable and Accelerated Genome Analysis

The diverse demands of genomic applications require customized hardware solutions to optimize performance and energy efficiency. We co-design hardware, algorithms, and software for end-to-end genome analysis by exploiting emerging technologies, minimizing data movement within the system, and targeting low-power edge platforms for real-time in-the-field use.

Algorithms and AI for Sequence Analysis

Accurate and scalable sequence analysis underpins many applications in computational genomics. This direction aims to design algorithms and AI methods that 1) tolerate noise and variation beyond exact matching, 2) eliminate redundant computation as references and datasets evolve, and 3) enable population-scale analysis with low latency and energy.

Multimodal Systems for Biological Data Analysis — Key Publications

Hardware-Algorithm-Software Co-Design — Key Publications

Algorithms and AI for Sequence Analysis — Key Publications

  • BLEND: a fast, memory-efficient and accurate mechanism to find fuzzy seed matches in genome analysis
    Can Firtina, Jisung Park, Mohammed Alser, Jeremie S Kim, Damla Senol Cali, Taha Shahroodi, Nika Mansouri Ghiasi, Gagandeep Singh, Konstantinos Kanellopoulos, Can Alkan, and Onur Mutlu
    NAR Genomics and Bioinformatics (NARGAB), March 2023
    [online link] [pdf] [code]
    Preprint (bioRxiv) [online link] [pdf]
    Social Media Thread [Twitter (X)] [Twitter (X)]
  • Apollo: a sequencing-technology-independent, scalable and accurate assembly polishing algorithm
    Can Firtina, Jeremie S. Kim, Mohammed Alser, Damla Senol Cali, A Ercument Cicek, Can Alkan, and Onur Mutlu
    Bioinformatics, June 2020
    [online link] [pdf] [code]
  • TargetCall: Eliminating the Wasted Computation in Basecalling via Pre-Basecalling Filtering
    Meryem Banu Cavlak, Gagandeep Singh, Mohammed Alser, Can Firtina, Joel Lindegger, Mohammad Sadrosadati, Nika Mansouri Ghiasi, Can Alkan, and Onur Mutlu
    Frontiers in Genetics, September 2024
    [online link] [pdf] [code]
  • RUBICON: a framework for designing efficient deep learning-based genomic basecallers
    Gagandeep Singh, Mohammed Alser, Kristof Denolf, Can Firtina, Alireza Khodamoradi, Meryem Banu Cavlak, Henk Corporaal, and Onur Mutlu
    Genome Biology, February 2024
    [online link] [pdf] [code]