Algorithms for the analysis of biological sequence data

About this book

Natural mass spectra are noisy with typically only 10-20% of total spectra yielding a high confidence peptide determination. In the first part of my dissertation, I show that by using a simple algorithm, it is possible to predict a high percentage of the predominant peaks in naturally occurring spectra. These predictions can provide intensity information which can be used to validate spectra or to filter spectra prior to de novo sequencing. In addition to spectral prediction, I describe the MDQUEST algorithm for mass spectra interpretation, the DREDGE algorithm which addresses the complex mapping of peptides to proteins, and the Fix Mix method for label free protein quantification. The second part of my dissertation focuses on computational algorithms for analysis of motifs surrounding post-translational modifications. I developed the motif-x and scan-x web-based tools to allow researchers to find motifs in their own data sets. I describe how motif-x and scan-x can be used on very large-scale data sets for prediction of post-translational modifications in entire proteomes. Using 51,808 known phosphorylation and acetylation modification sites from large-scale data sets, we first determined all significant motifs. These motifs were then used to scan entire proteomes to predict, with 99% specificity, an additional 30,035 phosphorylated residues in man, 18,941 in mouse, 8,936 in drosophila and 4,697 in yeast; and 18,392 acetylated residues in man. I also discuss applications of motif finding with motif-x and scan-x using a number of examples to illustrate their utility to create biological insights and to direct new experimentation. Proteomics and genomics are inextricably related, and in the final part of my research, I developed an algorithm called MIPTAG Pro to design thousands of molecular inversion probes for targeted genomic sequencing for any genomic targets with high coverage--even in repetitive regions of the genome. This probe design and capture technology is applied to sequence particular target genes and loci that are relevant to disease and also for sequencing the entire human exome.