Compression algorithms for biomedical signals and nanopore sequencing data
The massive generation of biological digital information creates various computing challenges such as its storage and transmission. For example, biomedical signals, such as electroencephalograms (EEG), are recorded by multiple sensors over long periods of time, resulting in large volumes of data. An...
Saved in:
| Main Author: | |
|---|---|
| Format: | doctoralThesis |
| Language: | English |
| Published: |
2021
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/20.500.12008/33017 |
| Tags: |
No Tags, Be the first to tag this record!
|
| Summary: | The massive generation of biological digital information creates various computing challenges such as its storage and transmission. For example, biomedical signals, such as electroencephalograms (EEG), are recorded by multiple sensors over long periods of time, resulting in large volumes of data. Another example is genome DNA sequencing data, where the amount of data generated globally is seeing explosive growth, leading to increasing needs for processing, storage, and transmission resources. In this thesis we investigate the use of data compression techniques for this problem, in two different scenarios where computational efficiency is crucial. First we study the compression of multi-channel biomedical signals. We present a new lossless data compressor for multi-channel signals, GSC, which achieves compression performance similar to the state of the art, while being more computationally efficient than other available alternatives. The compressor uses two novel integer-based implementations of the predictive coding and expert advice schemes for multi-channel signals. We also develop a version of GSC optimized for EEG data. This version manages to significantly lower compression times while attaining similar compression performance for that specic type of signal. In a second scenario we study the compression of DNA sequencing data produced by nanopore sequencing technologies. We present two novel lossless compression algorithms specifically tailored to nanopore FASTQ files. ENANO is a reference-free compressor, which mainly focuses on the compression of quality scores. It achieves state of the art compression performance, while being fast and with low memory consumption when compared to other popular FASTQ compression tools. On the other hand, RENANO is a reference-based compressor, which improves on ENANO, by providing a more efficient base call sequence compression component. For RENANO two algorithms are introduced, corresponding to the following scenarios: a reference genome is available without cost to both the compressor and the decompressor; and the reference genome is available only on the compressor side, and a compacted version of the reference is included in the compressed le. Both algorithms of RENANO significantly improve the compression performance of ENANO, with similar compression times, and higher memory requirements. |
|---|