Practical Bioinformatics Mark Voorhies 5/26/2015 Mark Voorhies - PowerPoint PPT Presentation

Practical Bioinformatics Mark Voorhies 5/26/2015 Mark Voorhies Practical Bioinformatics

Habits are things you get for free, without requiring any special work. –Cory Doctorow Advice to Writers, 4/5/2012 Mark Voorhies Practical Bioinformatics

Why compare sequences? Mark Voorhies Practical Bioinformatics

EM: Training an HMM Mark Voorhies Practical Bioinformatics

EM: Estimating transcript abundances Constrain Update parameters Online EM estimated counts algorithm A C G T A C + G T Error probabilities Bias Output Input ∝ λ ∝ α L Targets (i −1) c m i =m i c −1 i–1 Capture target Fragment and m � sequences sequence i Get next read pair Update masses Relative Estimated Effective abundances counts counts P (−) P (−) P (−) L Align to target references P ( ) ∝ λ L · ρ · ω p |−,L p Augmented · φ − | p, − ,L alignment file Calculate assignment probabilities Roberts and Pachter, Nature Methods 10:71 Mark Voorhies Practical Bioinformatics

Evolution implies a self-consistent model Distances Topology (Pairwise relationships) (Evolutionary history) Mark Voorhies Practical Bioinformatics

Measure all pairwise distances by dynamic programming Mark Voorhies Practical Bioinformatics

Measure all pairwise distances by dynamic programming Mark Voorhies Practical Bioinformatics

Generate a guide tree by UPGMA Mark Voorhies Practical Bioinformatics

Generate a guide tree by UPGMA Mark Voorhies Practical Bioinformatics

Generate a guide tree by UPGMA Mark Voorhies Practical Bioinformatics

Generate a guide tree by UPGMA Mark Voorhies Practical Bioinformatics

Generate a guide tree by UPGMA Mark Voorhies Practical Bioinformatics

Progressive alignment following the guide tree Mark Voorhies Practical Bioinformatics

Progressive alignment following the guide tree Mark Voorhies Practical Bioinformatics

Progressive alignment following the guide tree Mark Voorhies Practical Bioinformatics

Measure distances directly from the alignment Mark Voorhies Practical Bioinformatics

Generate neighbor-joining tree from new distances Mark Voorhies Practical Bioinformatics

Generate neighbor-joining tree from new distances Mark Voorhies Practical Bioinformatics

Generate neighbor-joining tree from new distances Mark Voorhies Practical Bioinformatics

Generate bootstrap values from subsets of the alignment Mark Voorhies Practical Bioinformatics

Generating a multiple alignment in CLUSTALX Mark Voorhies Practical Bioinformatics

Generating a multiple alignment in CLUSTALX Mark Voorhies Practical Bioinformatics

Generating a neighbor joining tree in CLUSTALX Mark Voorhies Practical Bioinformatics

Viewing the alignment and tree in JALVIEW Mark Voorhies Practical Bioinformatics

Related tools Protein Multiple Alignment MUSCLE Clustal Omega Probcons hmmalign (HMMer3) Tree Building MrBayes (Bayesian MCMC) PhyML (maximum likelihood) RaxML (fast maximum likelihood) FastTree2 (very large heuristic trees) Mark Voorhies Practical Bioinformatics

Homework Finish your dynamic programming implementation. Mark Voorhies Practical Bioinformatics