HMMER tutorial

1
HMMER tutorial

• ???
• g39208007_at_ym.edu.tw

2
Account

• IP 140.129.78.120
• Account binfo2005
• Password 2005binfo

3
HMMER

• http//hmmer.wustl.edu/
• The theory behind profile HMMs R. Durbin, S.
  Eddy, A. Krogh, and G. Mitchison, Biological
  sequence analysis probabilistic models of
  proteins and nucleic acids, Cambridge University
  Press, 1998.

4
Flowchart
http//bioweb.pasteur.fr/seqanal/motif/hmmer-uk.ht
ml
5
Format of input alignment files

• Output of CLUSTAL family of programs
• Wisconsin/GCG MSF format
• the input format for the PHYLIP phylogenetic
  analysis programs
• aligned FASTA format
• Stockholm format (HMMERs native format, used by
  the Pfam and Rfam databases)
• SELEX format

6
Searching a sequence database with a single
profile HMM

• build a profile HMM with hmmbuildgt hmmbuild
  globin.hmm globins50.msf
• calibrate the profile HMM with hmmcalibrategt
  hmmcalibrate globin.hmm
• search the sequence database with hmmsearchgt
  hmmsearch globin.hmm Artemia.fa

7
(No Transcript)
8
(No Transcript)
9
local alignment versus global alignment

• To HMMER, whether local or global alignments are
  allowed is part of the model, rather than being
  accomplished by running a different algorithm.
• you need to choose what kind of alignments you
  want to allow when you build the model with
  hmmbuild.
• By default, hmmbuild builds models which allow
  alignments that are global with respect
• to the HMM, local with respect to the sequence,
  and allows multiple domains to hit per sequence.

10
Searching a query sequence against a profile HMM
database

• creating your own profile HMM databasegt hmmbuild
  -A myhmms rrm.stogt hmmbuild -A myhmms fn3.stogt
  hmmbuild -A myhmms pkinase.stogt hmmcalibrate
  myhmms
• parsing the domain structure of a sequence with
  hmmpfamgt hmmpfam myhmms 7LES DROME

11
Creating and maintaining multiple alignments with
hmmalign

• Another use of profile HMMs is to create multiple
  sequence alignments of large numbers of
  sequences.
• A profile HMM can be build of a seed alignment
  of a small number of representative sequences,
  and this profile HMM can be used to efficiently
  align any number of additional sequences.
• gt hmmalign -o globins630.ali globin.hmm
  globins630.fa

12
(No Transcript)
13
(No Transcript)
14
HMMER scoring and determining significance

• HMMER gives you at least two scoring criteria to
  judge by the HMMER raw score, and an E-value.
• The E-value is calculated from the bit score. It
  tells you how many false positives you would have
  expected to see at or above this bit score.
• HMMER bit scores reflect whether the sequence is
  a better match to the profile model (positive
  score) or to the null model of nonhomologous
  sequences (negative score).

15
hmmsearch output
16
(No Transcript)
17
(No Transcript)
18

• Building a model
• hmmbuild From a multiple sequence alignment
• Using a model
• hmmalign Align sequences to an existing model
  (outputs a multiple alignment)
• hmmconvert Convert a model into different formats
  
• hmmcalibrate Takes an HMM and empirically
  determines parameters that are used to make
  searches more sensitive, by calculating more
  accurate expectation value scores (E-values)
• hmmemit Emit sequences probabilistically from a
  profile HMM
• hmmsearch Search a sequence database for matches
  to an HMM
• HMMs Databases
• hmmfetch Get a single model from an HMM database
• hmmindex Index an HMM database (not available on
  the WEB server)
• hmmpfam Search an HMM database for matches to a
  query sequence
• Other programs
• alistat Show some simple statistics about a
  sequence alignment file
• seqstat Show some simple statistics about a
  sequence file
• getseq Retrieve a (sub-)sequence from a sequence
  file (not available on the WEB server)
• sreformat Reformat a sequence(s) or alignment
  file into a different format

19
References

• HMMER user guide
• Eddy SR. (1998) Profile hidden Markov models.
  Bioinformatics.

20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
Related links

• HMMER http//hmmer.wustl.edu/
• SAM http//www.cse.ucsc.edu/research/compbio/sam.h
  tml
• PFTOOLS http//www.isrec.isb-sib.ch/ftp-server/pft
  ools/
• HMMpro http//www.netid.com/html/hmmpro.html
• GENEWISE http//www.ebi.ac.uk/Wise2/
• PROBE ftp//ftp.ncbi.nih.gov/pub/neuwald/probe1.0/
  
• META-MEME http//metameme.sdsc.edu/
• BLOCKS http//www.blocks.fhcrc.org/
• PSI-BLAST http//www.ncbi.nlm.nih.gov/BLAST/newbla
  st.html

27
Homework Search for homologies with hidden
Markov models

• Obtain the UniProtKB/Swiss-Prot entry P10242 of
  the myb proto-oncogene protein (AC P10242, entry
  MYB_HUMAN)
• Take the amino acid sequence of the myb protein
  and search against the NCBI nr protein database
  with BLASTp to obtain a HMM for myb-domains and
  use this HMM for searching against the
  UniProt-SwissProt protein database.
• Select 10 myb-domains while screening the hits of
  the BLASTp search and copy the corresponding
  parts of the sequences to a file in fasta-format
• Do a multiple sequence alignment with these ten
  myb-domains by ClustalW.

28
Homework Search for homologies with hidden
Markov models (cont.)

• Download HMMER from http//hmmer.wustl.edu/ and
  install.
• Build and calibrate a HMM of these myb-domains by
  means of hmmbuild and hmmcalibrate.
• Use hmmsearch to search against the
  UniProt-SwissProt protein library with the HMM of
  the myb-domains.
• Screen the hits, build a new HMM including
  selected hits and hmmsearch again.
• How many hits do you get? What are they?

29
HMM
30
Some examples