Machine Learning. HMM applications in computational biology

Size: px

Start display at page:

Download "Machine Learning. HMM applications in computational biology"

Aubrey McDaniel
5 years ago
Views:

1 Machine Learning HMM applications in computational biology

2 Central dogma DNA CCTGAGCCAACTATTGATGAA transcription mrna CCUGAGCCAACUAUUGAUGAA translation Protein PEPTIDE 2

3 Biological data is rapidly accumulating Transcription factors Next generation sequencing DNA transcription RNA translation Proteins

4 DNA Biological data is rapidly Transcription factors transcription accumulating Array / sequencing technology RNA translation Proteins

5 Biological data is rapidly accumulating Transcription factors Protein interactions DNA transcription RNA translation Proteins 38,000 identified interactions Hundreds of thousands of predictions

6 6

FDA Approves Gene-Based Breast Cancer MammaPrint is a DNA microarray-based test that measures the activity of 70 genes in a sample of a woman's breast-cancer tumor and

7 FDA Approves Gene-Based Breast Cancer MammaPrint is a DNA microarray-based test that measures the activity of 70 genes in a sample of a woman's breast-cancer tumor and then uses a specific formula to determine whether the patient is deemed low risk or high risk for the spread of the cancer to another site. Test* *Washington Post, 2/06/2007

8 8

9 Active Learning 9

10 Sequencing DNA First human genome draft in 2001 Due to accumulated errors, we could only reliably read at most nucleotides.

11 Shotgun Sequencing Wikipedia

12 Caveats Errors in reading Non-trivial assembly task: repeats in the genome MacCallum et al., GB 2009

13 Error Correction in DNA sequencing The fragmentation happens at random locations of the molecules. We expect all positions in the genome to have the same # number of reads K-mers = substrings of length K of the reads. Errors create error k-mers. Kellly et al., GB 2010

14 Transcriptome Shotgun Sequencing Miescher Laboratory Sequencing RNA molecule transcripts. Reminder: (mrna) Transcripts are expression products of genes. Different genes having different expression levels so some transcripts are more or less abundant than others.

15 Challenges Large datasets: millions reads of bps. Memory efficiency: Too time consuming to perform outmemory processing of data. DNA Sequencing + others : alternative splicing, RNA editing, post-transcription modification.

16 Errors are non uniformly distributed Some transcripts are more prone to errors Errors are harder to correct in reads from lowly expressed transcripts

17 SEECER Error Correction + Consensus sequence estimation for RNA-Seq data

18 Key idea: HMM model Salmela et al., Bioinformatics 2011 The way sequencers work: Read letter by letter sequentially Possible errors: Insertion, Deletion or Misread of a nucleotide

20 Building (Learning) the HMMs and Making Corrections (Inference) Learning = Expectation-Maximization Inference = Viterbi algorithm Seeding: Guessing possible reads using k-mer overlaps. Constructing the HMM from these reads. Speed up: The k-mer overlaps yield approximate multiple alignments of reads. We can learn HMM parameters from this directly.

21 Clustering to improve seeding Real biological differences should be supported by a set of reads with similar mismatches to the consensus

1. Clustering positions with mismatches to

Use Spectral clustering to find clusters of

22 1. Clustering positions with mismatches to identify clusters of correlated positions. 2. Build a similarity matrix between these positions. 3. Use Spectral clustering to find clusters of correlated positions. 4. Filter reads have mismatches in these clusters.

23 Comparison to other methods

24 Using the corrected reads, the assembler can recover more transcripts

25 Analysis of sea cucumber data B

26 Data integration in biology

27 Key problem: Most high-throughput data is static Time-series measurements Static data sources Sequencing motif CHIP-chip microarray PPI Time

28 DREM: Dynamic Regulatory Events Miner

29 a Expression Level Time Series Expression Data b TF A Static TF-DNA Binding Data time TF B TF C TF D c Expression Level Model Structure d IOHMM Model 0.1? time ?

30 Things are a bit more complicated: Real data

31 A Hidden Markov Model T t t t n i T t t t i H i H p i H i O p O H L )) ( ) ( ( )) ( ) ( ( ) ;, ( Hidden States Observed outputs (expression levels) t=0 t=1 t=2 t=3 H 0 H 1 H 2 H 3 O 0 O 1 O 2 O 3 Schliep et al Bioinformatics

32 Input Output Hidden Markov Model Input (Static TF-gene interactions) I g Hidden States (transitions between states form a tree structure) H 0 H 1 H 2 H 3 Emissions (Distribution of expression values) Log Likelihood O 0 O 1 O 2 O 3 t=0 t=1 t=2 t=3 Sum over all genes Sum over all paths Q Product over all Gaussian emission density values on path Product over all transition probabilities on path

33 E. coli. response Stem cells differentiation PLoS Comp. Bio Nature MSB 2011 Mouse Immune response Fly development Science 2010 IRF7 Genome Research 2010, PLoS ONE 2011

34 Things that work Approximate learning to speed up on large datasets. In real world, one technique is not enough. A solution involves using many techniques. Precision and Recall are trade-offs.

Computational Genomics. Reconstructing signaling and dynamic regulatory networks

Computational Genomics. Reconstructing signaling and dynamic regulatory networks 02-710 Computational Genomics Reconstructing signaling and dynamic regulatory networks Input Output Hidden Markov Model Input (Static transcription factorgene interactions) Bengio and Frasconi, NIPS 1995