14 - Genomics
from Part V - Applications
Published online by Cambridge University Press: 05 May 2015
Summary
We shall now explore how the techniques from the previous chapters can be used in studying the genome of a species. We assume here that we have available the so-called finalized genome assembly of a species. Ideally, this means the complete sequences of each chromosome in a species. However, in practice, this consists of chromosome sequences containing large unknown substrings, and some unmapped contigs/scaffolds.
To fix our mindset, we assume that the species under consideration is a diploid organism (like a human). We also assume that the assembled genome is concatenated into one long sequence T with some necessary markers added to separate its contents, and with some auxiliary bookkeeping data structures helpful for mapping back to the chromosome representation.
We start with a peak detection problem in Insight 14.1 that gives a direct connection to the segmentation problem motivating our study of hidden Markov models in Chapter 7. Then we proceed into variation calling, where we mostly formulate new problems assuming read alignment as input. The results of variant calling and read alignments are then given as the input to haplotype assembly.
Insight 14.1 Peak detection and HMMs
The coverage profile of the genome is an array storing for each position the amount of reads aligned to cover that position. In ChIP-sequencing and bisulfite sequencing only some parts of the genome should be covered by reads, so that clear peak areas should be noticeable in the coverage profile. In targeted resequencing, the areas of interest are known beforehand, so automatic detection of peak areas is not that relevant (although usually the targeting is not quite that accurate). Peak detection from a signal is a classical signal processing task, so there are many existing generalpurpose peak detectors available. In order to use the machinery we developed earlier, let us consider how HMMs could be used for our peak detection task.
[…]
- Type
- Chapter
- Information
- Genome-Scale Algorithm DesignBiological Sequence Analysis in the Era of High-Throughput Sequencing, pp. 307 - 324Publisher: Cambridge University PressPrint publication year: 2015