# Tutorial 8 - Hidden Markov models II.

## Problem 1 - Profile HMM construction

Consider the following multiple sequence alignmet. $$\begin{array} \mathrm{A} & \mathrm{C} & \mathrm{D} & \mathrm{E} & \mathrm{F} & \mathrm{A} & \mathrm{C} & \mathrm{A} & \mathrm{F} \\ \mathrm{A} & \mathrm{F} & \mathrm{D} & \mathrm{A} & \mathrm{\_} & \mathrm{\_} & \mathrm{\_} & \mathrm{C} & \mathrm{F} \\ \mathrm{A} & \mathrm{\_} & \mathrm{\_} & \mathrm{E} & \mathrm{F} & \mathrm{D} & \mathrm{\_} & \mathrm{F} & \mathrm{C} \\ \mathrm{A} & \mathrm{C} & \mathrm{C} & \mathrm{E} & \mathrm{F} & \mathrm{\_} & \mathrm{\_} & \mathrm{A} & \mathrm{C} \\ \mathrm{A} & \mathrm{D} & \mathrm{D} & \mathrm{E} & \mathrm{F} & \mathrm{A} & \mathrm{A} & \mathrm{A} & \mathrm{F} \\ \end{array}$$ Use threshold $\theta = 2$ to ignore columns which have too many gaps. Construct an HMM profile based on this multiple sequence alignment.

Imagine that you have found a new protein located in human heart muscle and some other strong muscles in the body. You believe that the protein has some connection with the ability of muscles to produce energy. Use either the Pfam database or the EBI HMMER search tool to identify the protein family to which the protein belongs.

The protein sequence is below.

MGLSDGEWQLVLNVWGKVEADIPGHGQEVLIRLFKGHPETLEKFDKFKHLKSEDEMKASE
DLKKHGATVLTALGGILKKKGHHEAEIKPLAQSHATKHKIPVKYLEFISECIIQVLQSKH
PGDFGADAQGAMNKALELFRKDMASNYKELGFQG

You may find the following alignment. Answer the questions below.

Which protein family does your protein belong to? What is this family responsible for/what is its purpose?

The quality of search is quantified in so-called $E$-value (sometimes Expect value). Find what this quantity means. What is the difference between the $E$-value and $p$-value? Based on the reported $E$-value, do you believe that the protein family is the same as the one reported by the database search?

The protein is, in fact, not new. Use the UniProt BLAST tool to find the true name of the protein. Knowing the protein name, answer the following questions:

• What is the purpose of the protein?
• Do we need this protein, i.e., is a malfunction of this protein lethal?
• Why it is good for organisms to have several proteins capable of the same thing? How do we call two proteins with the same origin?

This protein was not chosen arbitrarily. There was a Nobel prize awarded for research of this protein in 1962. Who received the prize and why? You will talk more about this field in the second part of the course.

## Problem 3 - exon prediction

There are many hidden Markov models tools capable of predicting introns and exons. Their power is in understanding the semantics of genes, and they provide, therefore, a good accuracy. Consider the sequence below (you can also download it as a FASTA file).

>NG_008301.1:5001-8050 Homo sapiens keratin 16 (KRT16), RefSeqGene on chromosome 17
AGTTAGGAGGGCCCCGCCTTCCCCAGCTGCATATAAAGGTCTCTGGGGTTGGAGGCAGCCACAGCACGCT
CTCAGCCTTCCTGAGCACCTTTCCTTCTTTCAGCCAACTGCTCACTCGCTCACCTCCCTCCTTGGCACCA
TGACCACCTGCAGCCGCCAGTTCACCTCCTCCAGCTCCATGAAGGGCTCCTGCGGCATCGGAGGCGGCAT
CGGGGGCGGCTCCAGCCGCATCTCCTCCGTCCTGGCCGGAGGGTCCTGCCGTGCCCCCAGCACCTACGGG
GGCGGCCTGTCTGTCTCCTCTCGCTTCTCCTCTGGGGGAGCCTGCGGGCTGGGGGGCGGCTATGGCGGTG
GCTTCAGCAGCAGCAGCAGCTTTGGTAGTGGCTTCGGGGGAGGATATGGTGGTGGCCTTGGTGCTGGCTT
CGGTGGTGGCTTGGGTGCTGGCTTTGGTGGTGGTTTTGCTGGTGGTGATGGGCTTCTGGTGGGCAGTGAG
AAGGTGACCATGCAGAACCTCAATGACCGCCTGGCCTCCTACCTGGACAAGGTGCGTGCTCTGGAGGAGG
CCAACGCCGACCTGGAAGTGAAGATCCGTGACTGGTACCAGAGGCAGCGGCCCAGTGAGATCAAAGACTA
CAGTCCCTACTTCAAGACCATCGAGGACCTGAGGAACAAGGTGGGTGACTTTGGTGTATGGAGCACTGAG
AGAGGCTGGGGCTACAGTGGCCCTTGGGATACCTCTTTTTAGCAATTACACTTTACAAACAGGGAGACTG
GGCACCTTTGGGGAGTGGCCAGGATCACCCAGGGAAGTGGTAGCAGAGGGTCCCTTTTCAGTATCTCTGT
GCCCGGACTGGGGCTGTTACCCTAAATCTCTTATTTCCTTCAAGGGTTCAGCTGCAAGTTCAGCTTCCCT
GCCTTGGGCCCAGGAAGGGGGTGATCGGGATGGAGTGCATCCCTACGTAGCCTGAGCTGGTGGAGAAGGC
ATGCCAGCCCTGCCAGCCAGAAGACTTCCAGATTTGGGGCGGTTCCTTTTGCCCCTTTCTGCCTTTCATG
CTCAAGTAGTAAGGTCCTTGGCTGACCAGGGCTCCTGTCCTCCATCCCCACTCCAGATCATTGCGGCCAC
CATTGAGAATGCGCAGCCCATTTTGCAGATTGACAATGCCAGGCTGGCAGCCGATGACTTCAGGACCAAG
TGAGCAGCCAGCATGGTGGGCTGGGGGCAGAGGGCAAGGGACAAAGAGTGGGGCGGTCCACCCAGCAGGG
CCAGCAGACCCCGAGCCTCAGAATCCTCAGGGCTGCAGCCTGAGGACCTGACCTCTGTCCTGCCAGGTAT
GAGCATGAACTGGCCCTGCGGCAGACTGTGGAGGCCGACGTCAATGGCCTGCGCCGGGTGTTGGATGAGC
TGACCCTGGCCAGGACTGACCTGGAGATGCAGATCGAAGGCCTGAAGGAGGAGCTGGCCTACCTGAGGAA
GAACCACGAGGAGGTACGGTCGCTGCTGGCTTCCGGGGTGGGAGGCTGGTTTGGTGGGGTTGCCAGATGC
ACCCAGGGCCAGGAGAGAAGTCTGCTGAACTGACCGCCTCCTGCCATCCCTTCCCAGGAGATGCTTGCTC
TGAGAGGTCAGACCGGCGGAGATGTGAACGTGGAGATGGATGCTGCACCTGGCGTGGACCTGAGCCGCAT
CCTGAATGAGATGCGTGACCAGTACGAGCAGATGGCAGAGAAAAACCGCAGAGACGCTGAGACCTGGTTC
CTGAGCAAGGTGGGGCTCGGGCCCGCAGTGAGCCTGCAGCACTTCCCAGCTGGGGGCTTTGGGAGAGCCT
CACCTTTCACTCTGCTTTCCTGCCTCAGACCGAGGAGCTGAACAAAGAAGTGGCCTCCAACAGCGAACTG
GTACAGAGCAGCCGCAGTGAGGTGACGGAGCTCCGGAGGGTGCTCCAGGGCCTGGAGATTGAGCTGCAGT
CCCAGCTCAGCATGGTATGAAGGACCCAGCACAGCAGCAGCCCCCAAGTCACCAGTAATGGCCACCACCC
CCTCAAAAAGCCACAGTCTAGTTCCACCTTTCTTTTCTCAGGATGGGACCAGGGGACTCATGGGACCGTT
ATATAGATAGAGAAACTAAGCCCTAGAATAGTGGGCTAGCTTTTCTCCATATTGTCTGGCCCATCAGTAC
CCCAACTGGGATCAAAATCCAGGCATCTCTCAAAAAACATGCCCAGAGACCTGGAGGAACAGGAGTGACC
ACCTCCATGGACTCTTTTTCTCTCTCTCACTTGCAGAAAGCATCCCTGGAGAACAGCCTGGAGGAGACCA
AAGGCCGCTACTGCATGCAGCTGTCCCAGATCCAGGGACTGATTGGCAGTGTGGAGGAGCAGCTGGCCCA
GCTACGCTGTGAGATGGAGCAGCAGAGCCAGGAGTACCAGATCTTGCTGGATGTGAAGACGCGGCTGGAG
CAGGAGATTGCCACCTACCGCCGCCTGCTGGAGGGCGAGGATGCCCAGTGAGTCCCAGGCCCCTCAGTTC
TGCCTCCCAGACCCTTTAGCCCCCCTGCTGCTCTCAGCACAACTGACTGCCCTGCTTTTTCTCTCCCACA
GCCTTTCCTCCCAGCAAGCATCTGGCCAATCCTATTCTTCCCGCGAGGGTAAGGCTTCTGAGGCTCCCCG
GCACTGCAGCCCCTCTGCCTGTTTCCATGGAGTGGGGGCTGGGCCCTTCTCCTCAGAGCTCCCAGCCCTC
CCTTCTCCCTGCCCTGGAGTCAGCTTAGCTCTCAGACCCCTTCTCACCTCCTCTTCTCTCTCCCACAGTC
TTCACCTCCTCCTCGTCCTCTTCGAGCCGTCAGACCCGGCCCATCCTCAAGGAGCAGAGCTCATCCAGCT
TCAGCCAGGGCCAGAGCTCCTAGAACTGAGCTGCCTCTACCACAGCCTCCTGCCCACCAGCTGGCCTCAC
CTCCTGAAGGCCCGGGTCAGGACCCTGCTCTCCTGGCGCAGTTCCCAGCTATCTCCCCTGCTCCTCTGCT
GGTGGTGGGCTAATAAAGCTGACTTTCTGGTTGATGCAAA

The first tool we will use is HMMgene on http://www.cbs.dtu.dk/services/HMMgene/. Use this tool to classify introns and exons of the sequence above.

In the output, you may notice the program to claim that you inserted two sequences, even though you used only a single sequence. Why does the program report two of them?

## gff-version 1
## date: Thu Apr  4 10:57:50 2019
## HMMgene1.1e (human model sim10gc.C.bsmod)

# SEQ:  3050 (+) A:593 C:922 G:903 T:632
HMMgene1.1e	firstex	140	670	# ...
# ..........
# SEQ:  3050 (-) A:632 C:903 G:922 T:593
HMMgene1.1e	exon_1	1824	1941	0.293	-	1	# ...
# ..........

How many introns and exons were predicted?

So far we do not know anything about the accuracy of the result. We shall, therefore, use a second tool to get additional information. Use GENSCAN on http://genes.mit.edu/GENSCAN.html to predict introns and exons for the same sequence.

Now compare the results. From the sequence annotation, you may see that the protein is keratin, which is a base stone of hair, nails, and cytoskeleton. Use the NCBI nucleotide archive to search for the sequence and annotation.

Compare the predicted introns and exons with the official annotation. How many differences did you find? What are the differences between the predicted peptide sequence and the one in the NCBI database?

If you are interested in more details on how GENSCAN works, visit the last slides of this presentation.

## Assignment 4 - gene finding

Work individually on the fourth programming assignment. The deadline is on May 13, 2020. Upload in BRUTE.

courses/bin/tutorials/tutorial8.txt · Last modified: 2021/02/12 15:24 (external edit)