Prokaryotic Annotation

Overview

Michelle Gwinn
August, 2005
 Annotation

• dictionary definition of “to annotate”:

– “to make or furnish critical or explanatory notes or
comment”
• some of what this includes for genomics
– gene product names
– functional characteristics of gene products
– physical characteristics of gene/protein/genome
– overall metabolic profile of the organism
• elements of the annotation process
– gene finding
– homology searches
– functional assignment
– ORF management
– data availability
• manual vs. automatic
– computers do a fair job at preliminary annotation
– high quality annotation requires manual review
Finding Real
Genes
 ORFs vs. Genes
• ORF = open reading frame
– absence of translation “stop” codons (TAA, TAG, TGA)
• an ORF goes from “stop” to “stop”
– ORFs are found easily by one of many ORF
finding tools
– ORFs can easily occur by chance and since “stop”
codons are AT rich:
• GC rich DNA has, on average, more, longer ORFs
• AT rich DNA has, on average, fewer, shorter ORFs

• Gene
– requires translation “start” codon
• bacterial starts = ATG, GTG, TTG
• genes go from “start” to “stop”
– has biological significance
• catalytic or structural RNAs
• protein coding regions

• Telling the difference between random ORFs and

genes is the goal in the gene finding process.
 A DNA sequence has 6
possible translation frames

AT G CTTTG CTTG G ATG A G CTCATA start

TAC G A AAC G A AC CTACTC G A GTAT stop

Frame +1 codons = ATG CTT TG C TTG GAT GA G CTC ATA

M L C L D E L I
Frame +2 codons = TGC TTT GCT TG G ATG AG C TCA
M S S
Frame +3 codons = GCT TTG CTT GGA TGA GCT CAT
L L G *
Frame -1 codons = TAT G A G CTC ATC CAA GCA AAG CAT
Frame -2 codons = ATG AGC TCA TCC AAG CAA AGC
M S S S K Q S
Frame -3 codons = TGA GCT CAT CCA AGC AAA GCA
*
 6-Frame translations
In order to visualize the genes
within the context of their
neighbors along a DNA sequence,
the sequence is often represented
as a 6-frame translation. There
are 6 possible frames for
translation in every sequence of
DNA, 3 in the forward (+) direction
on the DNA sequence, and 3 in the
reverse (-) direction on the DNA
sequence. These are represented
as horizontal bars with vertical
marks for stops (long) and starts
(short) in the order shown below.

+3
+2
+1
-1
-2
-3

start

stop
These are some of the many ORFs in this graphic.
 Gene Finding with Glimmer

• Glimmer is a tool which uses Interpolated Markov

Models (IMMs) to predict which ORFs in a genome
are real genes.

• Glimmer does this by comparing the nucleotide

patterns of “known” real genes to the nucleotide
patterns of the ORFs in the whole genome. ORFs
with patterns similar to the real genes are considered
real themselves.

• Using Glimmer is a two-part process:

– Train Glimmer for the organism that was sequenced.
– Run the trained Glimmer against the genome
sequence.
 Gene finding with Glimmer:
Gathering the Training Set
Gather published sequences from
the organism sequenced

If you need more,

Find all ORFs

Two options to
get additional training genes

long ORFs (500-1000 ORFs with a significant

nucleotides depending on BLAST match to a
GC content of genome) protein from
that do not overlap another organism
each other (what we do at TIGR)

Training

IMM
built
computer algorithm in Glimmer
 Gene Finding with Glimmer
What happens during training.

Glimmer moves sequentially through each sequence

in the training set, recording the nucleotide that
occurs after each possible oligomer up to oligomers
of length eight
Example for a 5-mer:
ATGCGTAAGGCTTTCACAGTATGCGAGTAAGCTGCGTCGTAA
GG

ATGCGTAAGGCTTTCACAGTATGCGAGTAAGCTGCGTCGTAA
GG

ATGCGTAAGGCTTTCACAGTATGCGAGTAAGCTGCGTCGTAA
GG

ATGCGTAAGGCTTTCACAGTATGCGAGTAAGCTGCGTCGTAA
GG

Glimmer then calculates the statistical probability

ATGCGTAAGGCTTTCACAGTATGCGAGTAAGCTGCGTCGTAA
GG pattern appearing in a real gene. These
of each
probabilities form the statistical model of what a
real gene looks like in the given organism.

This model is then run against the complete

genome sequence. All ORFs above the chosen
minimum length (99 bp at TIGR) are scored
according to how closely they match the model
of a real gene.
 Candidate ORFs
6-frame translation map for a region of DNA
Stop codons (TAA, TAG, TGA) (long hash marks)
Start codons (ATG, GTG, TTG) (short hash marks)
+3

+2
+1
-1
-2
-3

= ORFs meeting minimum length = laterally transferred DNA

The coding sequencse resulting from the candidate ORFs are represented
by arrows, going from start to stop, the dotted line represents an ORF
with no start site. which therefore can not be a gene. A long ORF does
not necessarily result in a long putative gene.

Green ORFs scored well to the model, red ORFs scored less well. The
green ORFs are chosen by Glimmer as the set of likely genes and
numbered sequentially from the beginning of the DNA molecule on which
they reside. ORFs in the area of lateral transfer, although real genes,
often will not be chosen since they don’t match the model built from the
patterns of the genome as a whole. Often when viewing a 6-frame
translation, the genes are represented as arrows drawn above (or, as in
this slide, below) the 6-frame translation.

ORF00002 ORF00004

ORF00001 ORF00003
 Coordinates

Genes are mapped to the underlying genome sequence

via coordinates. Each gene is defined by two coordinates:
end5 (the 5 prime end of the gene) and end3 (the 3 prime
end of the gene).

Nucleotide #1 for each molecule in the genome is the

beginning of each final assembled molecule. Some
genomes have just one DNA molecule, some have
several (multiple chromosomes or plasmids).

0 10000

gene end5 end3

purple 12 527
red 802 675
blue 927 1543
green 9425 7894
pink 9575 9945

Note that for forward genes have end5<end3,

while reverse genes have end5>end3.
Determining How The
New Proteins Function
 Finding the function
of a new protein
• Experimental characterization
– mutant phenotype
– enzyme assay
– difficult on a whole-genome scale
• microarrays
– expression patterns
• large-scale mutant generation
– done in yeast

• Homology searching
– comparing sequences of unknown
function to those of known function
 Homology searching

• shared sequence implies shared

function
– binding sites
– catalytic sites
– full length match with significant identity
between amino acids (>35% minimum)
• but beware
– there are occurrences of proteins where one
amino acid substitution changes the function
of an enzyme
– all functional assignments made by
sequence similarity should be considered
putative until experimental characterization
confirms them
• identity vs. similarity
– identity means amino acids match exactly
– similarity means the amino acids share
similar structure and thus could carry out the
same or similar roles in the protein
 Protein Alignment Tools
• Local pairwise alignment tools do not worry
about matching proteins over their entire
lengths, they look for any regions of similarity
within the proteins that score well.
– BLAST
• fast
• comes in many varieties (see NCBI site)
– Smith-Waterman
• finds best out of all possible local alignments
• slow but sensitive

• Global pairwise alignment tools take two

sequences and attempt to find an alignment of
the two over their full lengths.
– Needleman-Wunsch
• finds best out of all possible alignments

• Multiple alignments are more meaningful than

pairwise alignments since it is much less likely
that several proteins will share sequence
similarity due to chance alone, than that 2 will
share sequence similarity due to chance alone.
Therefore, such shared similarity is more likely
to be indicative of shared function.
– HMMs
– motifs
 Sample Alignments

Pairwise

-two rows of amino acids compared to each other, the top row is the
search protein and the bottom row is the match protein, numbers
indicate amino acid position in the sequence
-solid lines between amino acids indicate identity
-dashed lines (colons) between amino acids indicate similarity

Multiple

Different shadings indicate amount of matching

 Useful Databases
Slide 1
• NCBI
– National Center for Biotechnology Information
– protein and DNA sequences
– taxonomy resource
– many other resources
• Omnium
– database that underlies TIGR’s CMR
– contains data from all completed sequenced bacterial
genomes
– data is downloaded from the sequencing center
• Enzyme Commission
– not sequence based
– catagorized collection of enzymatic reactions
– reactions have accession numbers indicating the type
of reaction
• Ex. 1.2.1.5
• KEGG, Metacyc, etc.
 Useful Databases
Slide 2
• Swiss-Prot
– European Bioinformatics Institute (EBI) and Swiss
Institute of Bioinformatics (SIB)
– all entries manually curated
– annotation includes
• links to references
• coordinates of protein features
• links to cross-referenced databases
– HMMs
– Enzyme Commission
• TrEMBL
– EBI and SIB
– entries have not been manually curated
– once they are accessions remain the same but move
into Swiss-Prot
• PIR (at Georgetown University)
• UniProt
– Swiss-Prot + TrEMBL + PIR
 NIAA

• Non-Identical Amino Acid

• TIGR’s protein file used for searching
• File composed of protein sequences
from several source databases
– Swiss-Prot
– Omnium
– NCBI
– PIR
• The file is made non-redundant
– identical protein sequences from the same
gene in the same organism that came into
the file from different source databases are
collapsed into one entry
– all of the protein’s accession numbers from
the various source databases where it is
found are stored linked to the protein
– users can always view the protein at the
source database
 NIAA entry

>biotin synthase, Escherichia coli

M A H RP R W TLS QVTELFEKPLLDLLFEA Q Q V H R Q HFDPR Q V Q VSTLLSIKTG A CPE D
P QSS RYKT GLEAERL MEVE QVLESARKAKAA GSTRFC M G AAW K N P H ER D M PYLE
KA M GLEAC MTLGTLSES Q A Q RLANA GLDYYN H NLDTSPEFY G NIITTRTY QE RLDTL
R DA GIKVCS G GIV GL GETVKD RA G LLLQLANLPTPPESVPIN M LVKVK G TPLAD N D D
FDFIRTIAVARIMM PTSYVRLSA G R E Q M N E QT Q A M C F MA G A N SIFYG C KLLTTPNPE
DL QLFRKLGLNPQ QTAVLA G D NE Q Q Q RLE Q AL MTPDTDEYYNAAAL

Source databases where this protein is found:

-Swiss-Prot, accession # SP:P12996

-Protein Information Resource, accession # PIR:JC2517
-NCBI’s GenBank, accession # GB:AAC73862.1

All of these are collapsed into one entry in NIAA that is linked
to all three accessions.
 Experimentally
Characterized Proteins
• It is important to know what proteins in
our search database are characterized.
– We store the accessions of proteins known
or suspected to be characterized in the
“characterized table” in our database
– A confidence status is assigned to each
entry in the characterized table.
• Annotators see this information in the
search results as color coded output:
– green = full experimental characterization
– red = automated process (Swiss-Prot parse)
– sky blue = partial characterization
– olive = trusted, used when multiple extremely good
lines of evidence exist for function but no experiment
has been done (rarely used)
– blue-green = fragment/domain has been characterized
– fuzzy gray = void, used to indicate that something that
was originally thought to be characterized really is not
(rarely used)
– gray = accession exists in the omnium only - therefore
represents automated annotation
• Our table does not contain all
characterized proteins, not even close.
– Any time additional characterized proteins
are found it is important that they be entered
into the table
 BLAST-extend-repraze (BER)
TIGR’s pairwise protein search tool
• Initial BLAST search
– against NIAA
– finds local alignments
– stores good hits in mini-database for each protein
• Protein sequence is extended by 300
nucleotides on each end and translated
(see later slide)
• A modified Smith-Waterman alignment is
generated with each sequence in the mini-
database
– extends local alignments as far as homology
continues over lengths of extended proteins
– produces a file of alignments between the query
protein and the match protein for each match
protein in the mini-database
– as the alignment generating algorithm builds the
alignment, if the level of similarity falls below the
necessary threshold, the program looks in
different frames and through stop codons for
homology to continue - this similarity can continue
into the upstream and/or downstream extensions
 BLAST-extend-repraze (BER)
genome.pep
vs. niaa
(non-identical amino acid)

BLAST

mini-db for mini-db for mini-db for mini-db for

protein #1 protein #2 protein #3 protein #3000

...
, ,
Significant hits put into mini-dbs for each protein
Extend 300 nucleotides
on both ends

modified Smith-
Waterman Alignment

vs.
extended protein
mini-database from BLAST search
 BER Alignment
An alignment like this will be generated for every match
protein in the mini-database. In the next slides we will
look closely at the types of information displayed here.
 BER Alignment detail:
Boxed Header

-The background color of this box will be gold if the protein is in

the characterized table and grey if it is not.

-The top bar lists the percent identity/similarity and the organism
from which the protein comes (if available).

-The bottom section lists all of the accession numbers and names
for all the instances of the match protein from the source
databases (used in building NIAA for the searches.)

-The accession numbers are links to pages for the match protein
in the source databases.

-A particular entry in the list will have colored text (the color
corresponding to its characterized status) if that is the accession
that is entered into the characterized table - this tells the
annotators which link they should follow to find experimental
characterization information. Only one accession for the match
protein need be in the characterized table for the header to turn
gold.

-There are links at the end of each line to enter the accession into
the characterized table or to edit an already existing entry in the
characterized table.
 BER Alignment detail:
alignment header

-It is most important to look at the range over which the alignment
stretches and the percent identity

-The top line show the amino acid coordinates over which the match
extends for our protein

-The second line shows the amino acid coordinates over which the
match extends for the match protein, along with the name and
accession of the match protein

-The last line indicates the number of amino acids in the alignment
found in each forward frame for the sequence as defined by the
coordinates of the gene. The primary frame is the one starting with
nucleotide one of the gene. If all is well with the protein, all of the
matching amino acids should be in frame 1.

-If there is a frameshift in the alignment (see later slide) the phrase
“Frame Shifts = #” will flash and indicate how many frameshifts there
are.
 BER Alignment detail:
alignment of amino acids

-In these alignments the codons of the DNA sequence read down in
columns with the corresponding amino acid underneath.

-The numbers refer to amino acid position. Position 1 is the first

amino acid of the protein. The first nucleotide of the codon coding
for amino acid 1 is nucleotide 1 of the coding sequence. Negative
amino acid numbers indicate positions upstream of the predicted
start of the protein.

-Vertical lines between amino acids of our protein and the match
protein (bottom line) indicate exact matches, dotted lines (colons)
indicate similar amino acids.

-Start sites are color coded: ATG is green, GTG is blue, TTG is
red/orange

-Stop codons are represented as asterisks in the amino acid

sequence. An open reading frame goes from an upstream stop
codon to the stop at the end of the protein, while the gene starts at
the chosen start codon.
 BER Skim
A list of best matches from niaa to the search protein with
statistics on length of match and BLAST p-value. Colored
backgrounds indicate presence in characterized table and
corresponding status.
 Extensions in BER
The extensions help in the detection of frameshifts (FS)
and point mutations resulting in in-frame stop codons
(PM). This is indicated when similarity extends outside
the coordinates of the protein coding sequence
end5 end3

300 bp ORFxxxxx 300 bp

search protein

match protein
normal full length match

! FS

similarity extending through and frameshift upstream or

downstream into extensions

* PM
similarity extending in the same frame through a stop codon
?
FS or PM ?
two functionally unrelated genes from other species matching
one of our proteins could indicate incorrectly fused ORFs
Frameshifted alignment
Hidden Markov Models - HMMs
• HMMs are statistical models of the
patterns of amino acids in a group of
functionally related proteins found
across species.
– this group is called the “seed”
– HMMs are built from multiple alignments of
the seed members.
• Proteins searched against an HMM
receive a score indicating how well
they match the model.
– Proteins scoring well to the model can be
expected to share the function that the HMM
represents.
• HMMs can be built at varying levels of
functional relationship.
– The most powerful level of relationship is
one representing the exact same function.
– It is important to know the kind of
relationship an HMM models to be able to
draw the correct conclusions from it
• Annotation can be attached to HMMs
– protein name
– gene symbol
– EC number
– role information
 TIGR’s HMM Isology Types
Equivalog: The supreme HMM, designed so that all members
and all proteins scoring above trusted share the same function.

Superfamily: This type of HMM describes a group of proteins

which have full length protein sequence similarity and have the
same domain architecture, but which do not necessarily have the
same function.

Subfamily: This type of HMM describes a group of proteins

which also have full length homology, which represent more
specific sub-groupings with a superfamily.

Domain: These HMMs describe a region of homology that is

not required to be the full length of a protein. The function of the
region may or may not be known.

Equivalog_domain: Describes a protein region with a

conserved function. It can be found as a single function protein or
part of a multifunctional protein.

Hypothetical_equivalog: These are built in the same way as

equivalogs except they are made from only conserved
hypothetical proteins. Therefore, although the function is not
known, it is believed that all proteins that score well to the HMM
share the same function.

Pfam: Indicates that no TIGR isology type has yet been

assigned to the Pfam HMM.
 Building HMMs
Collect proteins to be in the “seed”
(same function/similar domain/ family membership)

Generate and Curate Multiple Alignment of Seed proteins

Region of good alignment and closest similarity

Run HMM algorithm Search new model

Computes statistical this step may against all proteins
probabilities for amino acid need a few
patterns in the seed iterations

Choose “noise” and “trusted” cutoff scores based

on what scores the “known” vs. “unknown”
proteins receive. HMM is ready to go!
 Choosing cutoff scores
=search the new HMM against NIAA
=see the range of scores the match proteins receive
=do analysis to determine where known members score
=do analysis to determine where known non-members score
=set the cut-offs accordingly

matches (seed members bold) score

protein “definitely” 547
protein “absolutely” 501
protein “sure thing” 487
protein “confident” 398
protein “safe bet” 376
protein “very confident” 365
protein “has to be one” 355 250
protein “could be” 210
protein “maybe” 198
protein “not sure” 150 100
protein “no way” 74
protein “can’t be” 54
protein “not a chance” 47

=proteins that score above trusted can be considered part of the

protein family modeled by the HMM
=proteins that score below noise should not be considered part of
the protein family modeled by the HMM
=usefulness of an HMM is directly related to the care taken by the
person building the HMM since some steps are subjective
 HMM Searches

genome.pep
ATG_______________ HMM database
ATG__________
GTG________________________ vs.
ATG____________________ TIGRFAMS + Pfams
ATG_____________________________

HMM hits for HMM hits for HMM hits for

protein #1 protein #2 protein #3

TIGR01234 TIGR01004

, ,
PF00012

,
NONE
TIGR00005

etc.
Each protein in the genome is searched against all HMMs
in our db. Some will not have significant hits to any HMM,
some will have significant hits to several HMMs. Multiple
HMM hits can arise in many ways, for example: the same
protein could hit an equivalog model, a superfamily model
to which the equivalog function belongs, and a domain
model representing the catalytic domain for the particular
equivalog function. There is also overlap between TIGR
and Pfam HMMs.
 Evaluating HMM scores

Above trusted - protein is member of family HMM models

P
-50 0 100

N T

Below noise - protein is not member of family HMM models

-50 0 100

In-between noise and trusted - protein may be in family HMM models

-50 0 100

Above trusted with negative scores - protein is in family HMM models

-50 0 100
HMM Output in Manatee
 Genome Properties
• Used to get “the big picture” of an organism. Specifically to
record and/or predict the presence/absence of:
– metabolic pathways
• biotin biosynthesis
– cellular structures
• outer membrane
– traits
• anaerobic vs. aerobic
• optimal growth temperature
• Particular property has a given “state” in each organism, for
example:
– YES - the property is definitely present
– NO - the property is definitely not present
– Some evidence - the property may be present and more
investigation is required to make a determination
• The state of some properties can be determined
computationally
– metabolic pathway
• the property is defined be several reaction steps which are
modeled by HMMs
• HMM matches to steps in pathway indicate that the organism
has the property
• Other property’s states must be entered manually (growth
temp, anaerobic/aerobic, etc.)
• data for a particular genome viewable in Manatee
– links from HMM section
– links from gene list for role category
– entire list of properties and states can be viewed
• Searchable across genomes on the CMR site
– covered in the CMR segment of the course
“Biotin Biosynthesis”
Genome Property
“Cell Shape” Genome Property
 Paralogous Families
genome.pep
ATG_______________
genome.pep
ATG_______________
ATG__________ ATG__________
GTG________________________ vs. GTG________________________
ATG____________________ ATG____________________
ATG_____________________________ ATG_____________________________

Groups proteins from within the same genome into

families (minimum two members) according to
sequence similarity. First, proteins are clustered
according to HMM hits, second, other regions of the
proteins, not found in HMM hits, are searched and
clustered.
Reveals expansion/contraction of various
families of proteins in one genome verses
another.
Helps in annotation consistency,
frameshift detection, and start site editing.
Paralogous family
output in Manatee
 Other searches
• PROSITE Motifs
– collection of protein motifs associated with
active sites, binding sites, etc.
– help in classifying genes into functional
families when HMMs for that family have not
been built
• InterPro
– Brings together HMMs (both TIGR and
Pfam) Prosite motifs and other forms of
motif/domain clustering (Prints, Smart)
– Useful annotation information
– GO terms have been assigned to many of
these
• TmHMM
– an HMM that recognizes membrane spans
– a product of the Center for Biological
Sequence Analysis (CBS), Denmark
• Signal P
– potential secreted proteins
– another CBS product
• Lipoprotein
– potential lipoproteins
– this is actually a specific Prosite motif
 Other Searches/Information

• Molecular Weight/pI
• DNA repeats
• RNAs
– tRNAs are found using tRNAscan (Sean Eddy)
– structural RNAs are found using BLAST
searches
– We are starting to implement Rfam, a set of
HMMs modeling non-coding RNAs (Sanger,
WashU)
• GC content
– for the whole genome and individual genes
• terminators
• operons
 Making the annotations:
Assigning names and roles to the proteins
 Functional Assignments:
What we want to accomplish.

Name and associated info

Descriptive common name for the protein, with
as much specificity as the evidence supports;
gene symbol. EC number if protein is an
enzyme

Role
Both TIGR and Gene Ontology, to describe
what the protein is doing in the cell and why.

Supporting evidence:
HMMs, Prosite, InterPro
Characterized match from BER search
Paralogous family membership.
 Functional Assignments:
What we want to avoid!

Genome Rot!
Transitive Annotation: A is like
B, B is like C, C is like D, but A
is not like D

We take a very conservative

approach and err on the side of
missing homology rather than
stretching weak data.

Increasingly, the BER search

results are filled with sequences
from genome projects, the
names of those proteins can not
be considered reliable.
 AutoAnnotate

Software tool which gives a preliminary name

and role assignment to all the proteins in the
genome.

Makes decisions based on ranked evidence types

best
evidence
Equivalog HMM

Characterized BER match

Other HMMs

Other BER match

Etc.
 Manual Annotation:
Assigning Names to Proteins
 Functional Assignments:
High Confidence in Precise Function

Criteria:
-At least one good alignment (minimum 35% identity,
over the full lengths of both proteins) to a protein from
another organism that has been experimentally
characterized, preferably multiple such alignments.
-Above trusted cutoff hits to any HMMs for this gene.
-Conservation of active sites, binding sites, appropriate
number of membrane spans, etc.

Give the protein a specific name and

accompanying gene symbol, this is the only
confidence level where we assign gene
symbols. We default to E. coli gene symbols
when possible, for Gram positive genes we
use B. subtilis gene symbols.
Example:

name: “adenylosuccinate lyase”

gene symbol: purB
EC number: 4.3.2.2
 Functional Assignments:
High Confidence in Function,
Unsure of Specificity

A good example of this is seen with

transporters, what you’ll see:
-Multiple hits to a specific type of transporter
-Hits to appropriate HMMs
-The substrate identified for the proteins your protein
matches may not all be the same, but may fall into a
group, for example they are all sugars.

The name for a specific substrate:

“ribose ABC transporter, permease protein”

The name for specific function but a more general

substrate specificity:
“sugar ABC transporter, permease protein”

Sometimes it will not be possible to identify particular

substrate group, in that case:
“ABC transporter, permease protein”

Another example of known function but not exact

substrate:
“carbohydrate kinase, FGGY family”
 Functional Assignments:
Function Unclear

The “family” designation:

-No matches to specific characterized protein
-score above trusted cutoff to an HMM which defines a
family, but not a specific function.
“CbbY family protein”
The “homolog” designation:
-if match to a characterized protein is not good enough to
say for sure that the two proteins share function (in general,
less than 35% id)
-HMM match might be below trusted and above noise
-some active sites missing
OR
-good match to a function not expected in the organism
(like a photosynthesis gene in a non-photosynthetic bug)
“galactokinase homolog”
The “putative” designation is used when data is very
close to being enough for actual functional
assignment:
-has been largely replace by “homolog” and “family”
“putative galactokinase”
 Functional Assignments:
Hypotheticals

If a protein has no matches to any protein

from another species, HMM, Prosite, or
InterPro it is called:

“hypothetical protein”

If a “hypothetical protein” from one species

matches a “hypothetical protein” from
another, they both now become:

“conserved hypothetical protein”

 Functional Assignments:
Frameshifts and Point Mutations

Possible sequence errors detected in the BER

alignments are sent back to the lab for checking.

Sometimes an error in the sequence is found

and corrected.

In others the sequence is shown to be correct

and the protein is annotated to reflect the
presence of a disruption in the open reading
frame:

“great protein, authentic frameshift”

“fun protein, authentic point mutation”

 Functional Assignments:
Other ORF disruptions

-many FS/PM, ! ! !
“degenerate” ** *
-”truncation”

-”deletion”

-”insertion”
(~20-50aa)

-interruption
“interruption-N” N-term
(some genes)
C-term
“interruption-C”

-”fusion”

-”fragment”

These are given descriptive terms in the

common name and all are put into a “Disrupted
reading frame” role category to make them easy
to find.
 Manual Annotation:
Assigning Roles to Proteins
 TIGR Roles
TIGR bacterial roles were AutoAnnotate makes
first adapted from Monica a first pass at
Riley’s roles for E. coli - both assigning role, based
systems have since on roles associated
undergone much change. with HMMs or with
match proteins.
Unclassified (not a real role)
Amino acid biosynthesis
Purines, pyrimidines,
nucleosides, and nucleotides
Fatty acid and phospholipid
metabolism Human annotator
Biosynthesis of cofactors, checks and adjusts as
prosthetic groups, and carriers necessary.
Central intermediary metabolism
Energy metabolism
Transport and binding proteins
DNA metabolism
Transcription
Protein synthesis
Role Notes:
Protein Fate
Regulatory Functions Notes written by
Signal Transduction annotators expert in each
Cell envelope role category to aid other
Cellular processes annotators in knowing
Other categories what belongs in that
Unknown category and the TIGR
Hypothetical naming conventions for it.
Disrupted Reading Frame
 Gene Ontology (GO)
Consortium
• began as collaboration between the
databases for mouse, fly, and yeast, but
has grown considerably and TIGR is
now a member
• three controlled vocabularies for the
description of:
– molecular function (what a gene does)
– biological process (why a gene functions)
– cellular component (where a gene acts/lives)
• can reflect annotation with assignment
of GO terms
– GO terms exist at many levels of specifiicy
(granularity)
– assign GO terms with specificity appropriate for
what is known about the function of the protein in
question
– provides mechanism for storing evidence for GO
terms and thus annotation
• can be easily searched by a computer
and allows searches/comparisons
across species, kingdoms
• if everyone uses the same system, it
allows greater exchange of data
 GO term composition
• GO terms have 3 required parts
– ID number – unique stable ids
– Name
– Definition – the part of the term that the id
number actually refers to, if the name is changed
but the definition remains the same – the id stays
the same, but if the definition changes – a new
GO term must be made
• Other info connected to GO term
– comment
• gives additional information for proper
annotation, tells users why terms were
obsoleted
– cross reference
• ex. EC numbers
– synonyms
• alternate enzyme names
• abbreviations (TCA)
 Example GO term

•ID number: GO:0004076

•Name: biotin synthase activity
•Definition: Catalysis of the reaction\:
dethiobiotin + sulfur = biotin.
•comment: none
•cross reference: EC:2.8.1.6
•synonyms: none
•parent term: sulfurtransferase activity
(GO:0016783)
•relationship to parent: “is a”
 GO term relationships

• Each GO term has a relationship to at least

one other term
– process/function/component are roots
– terms always have at least one parent
– a term may have children and/or siblings, siblings
share the same parent
– as one moves down the tree from parents to
children, the functions, processes, and structures
become more specific (or granular) in nature
• GO is a DAG (directed acyclic graph)
– a term can have many parents (as opposed to a
hierarchical structure)
• relationship types
– is a (most terms)
• ribokinase “is a” kinase
– part of (generally found mostly in component)
• periplasm is “part of” a cell
– (regulates - arriving soon)
Example Tree
 Annotating with GO
• Decide what annotation the protein should have,
find the corresponding terms
– Your favorite tree viewing tool
• Manatee GO viewer
• AmiGO (on GO web page)
– Mapping files
• ec2go - a list of EC numbers and corresponding GO
terms
• Tigrfams2go - GO terms assigned to TIGR HMMs
– Search against proteins already annotated to GO
• GOst (at GO web page)
• GO correlations
• protein name search
• TIGR GO Blast
• Try to get a term from every ontology at the level of
specificity you are confidant of. Don’t be afraid to
use the “unknown” terms (there’s one in each
ontology).
• Assign as many terms as are appropriate to
completely describe what is known about the
protein (you can have multiple terms from each
ontology)
• Send annotation to GO to be placed in the
repository of annotated genes to be a resource to
the community
– currently 11 TIGR prokaryotic genomes at GO
 Functional confidence captured
with GO

Available evidence
for Function
catalytic activity
3 genes kinase activity
carbohydrate kinase activity
#1 ribokinase activity
-HMM for “ribokinase’ glucokinase activity
-characterized match to fructokinase activity
ribokinase

Process
#2
metabolism
-HMM for “kinase”
carbohydrate metabolism
-characterized matches
monosaccharide metabolism
to a “glucokinase’ AND a
hexose metabolism
‘fructokinase’
glucose metabolism
fructose metabolism
pentose metabolism
#3
ribose metabolism
-HMM for “kinase’
 GO Evidence
• Just as we store evidence for our annotation, we must
store evidence for GO term assignments:
– Assign Evidence Code
• Ev Codes tell users what kind of evidence was used
– sequence similarity (99% of our work) - ISS
– experimental characterization - IMP, IDA, etc.
• IEA - code for electronic annotation - immediately allows
users to tell manual curation from automatic
– Assign “Reference”
• PMID of paper describing characterization or method used
for annotation
• database reference (GO standards)
– Assign “with” (when appropriate)
• Used with ISS to store the accession of the thing the
sequence similarity is with
– Modifier column
• “contributes to” - use this modifier when you assign a
function term representing the function of a complex to
proteins that are part of the complex but can not
themselves carry out the function of the complex
• All accessions used as evidence must be represented
according to GO’s format – “database:accession”
(where “database” is the abbreviation defined at GO).
Manatee knows these rules and automatically puts
the accessions in the correct format.
– Examples
• TIGR_TIGRFAMS:TIGR01234
• Swiss-Prot:P12345
 GO Evidence codes

• IEA inferred from electronic annotation

• IC inferred by curator
• IDA inferred from direct assay - Enzyme assays
- In vitro reconstitution (e.g. transcription)
- Immunofluorescence (for cellular component)
- Cell fractionation (for cellular component)
- Physical interaction/binding
• IEP inferred from expression pattern
- Transcript levels (e.g. Northerns, microarray data)
- Protein levels (e.g. Western blots)
• IGI inferred from genetic interaction
- "Traditional" genetic interactions such as
suppressors, synthetic lethals, etc.
- Functional complementation
- Rescue experiments
- Inference about one gene drawn from the
phenotype of a mutation in a different gene.
• IMP inferred from mutant phenotype
- Any gene mutation/knockout
- Overexpression/ectopic expression of wild-type or mutant genes
- Anti-sense experiments
- RNAi experiments
- Specific protein inhibitors
• IPI inferred from physical interaction
- 2-hybrid interactions
- Co-purification
- Co-immunoprecipitation
- Ion/protein binding experiments
• ISS inferred from sequence or structural similarity
- Sequence similarity (homologue of/most closely related to)
- Recognized domains
- Structural similarity
- Southern blotting
• NAS non-traceable author statement • ND no biological data available
• TAS traceable author statement • NR not recorded

http://www.geneontology.org/GO.evidence.html
http://www.geneontology.org/GO.annotation.html
Association files at GO
 GO is a work in progress

• The GO actively requests user participation

in ontology development
– new terms
– changes to existing terms
– SourceForge site
• consortium meetings
• user meetings
• terms can become obsolete, but their ids
are never used again and they remain in the
ontolgies so people can track them
• as the ontologies change annotations must
be changed too - in particular annotations
to terms that have become obsolete (like
“toxin activity”)
ORF Management
and
Data Availability
 ORF management: Start site edits

What to consider:
- Start site frequency: ATG >> GTG >> TTG
- Ribosome Binding Site (RBS): a string of AG rich sequence
located 5-11 bp upstream of the start codon
- Similarity to match protein, both in BER and Paralogous
Family - probably the most important factor.
(Remember to note, that the DNA sequence reads down in
columns for each codon.)
-In the example below (showing just the beginning of one
BER alignment), homology starts exactly at the first atg (the
current chosen start, aa #1), there is a very favorable RBS
beginning 9bp upstream of this atg (gagggaga). There is no
reason to consider the ttg, and no justification for moving to
the second atg (this would cut off some similarity and it does
not have an RBS.)

RBS upstream This ORF’s

3 possible
of chosen start upstream boundary
start sites
 ORF management:
Overlaps and Intergenics

Overlap analysis
When two ORFs overlap (boxed areas), the one without similarity
to anything (another protein, an HMM, etc.) is removed. If both
don’t match anything, other considerations such as presence in a
putative operon and potential start codon quality are considered.
This process has both automated (for the easy ones) and manual
(for the hard ones) components. Small regions of overlap are
allowed (circle).

InterEvidence regions
Areas of the genome with no genes and areas within genes without
any kind of evidence (no match to another protein, HMM, etc., such
regions may include an entire gene in case of “hypothetical
proteins”) are translated in all 6 frames and searched against niaa.
Results are evaluated by the annotation team.
 Data Availability

• Publication
– TIGR staff/collaborators analysis of
genome data
• GenBank
– Sequence and annotation submitted
to GenBank at the time of publication
– Updates sent as needed
• Comprehensive Microbial
Resource (CMR)
– Data available for downloading
– extensive analyses within and
between genomes
 Useful links

• CMR Home
– http://www.tigr.org/tigr-
scripts/CMR2/CMRHomePage.spl
• SIB web site (Swiss-Prot, Prosite, etc.)
– http://www.expasy.org
• PIR
– http://pir.georgetown.edu
• NCBI
– http://www.ncbi.nlm.gov
• BLAST
– http://blast.wustl.edu
• GO
– http://www.geneontology.org
• TIGRFAM HMMs
– http://www.tigr.org/tigr-
scripts/CMR2/find_hmm.spl?db=CMR
– OR
– http://tigrblast.tigr.org/web-hmm/
 Acknowledgements

leading the effort: CMR team:

Owen White Tanja Davidsen (Team leader)
Jeremy Peterson Nikhat Zafar
Qi Yang
Prokaryotic Annotation
Bill Nelson (Team leader) HMM team:
Bob Dodson Dan Haft
Bob Deboy Jeremy Selengut
Scott Durkin
Sean Daugherty Building the tools:
Ramana Madupu Todd Creasy (head Manatee
Lauren Brinkac developer)
Steven Sullivan Liwei Zhou
M.J. Rosovitz Sam Angiuoli (and his team)
Sagar Kothari Anup Mahurkar (and his team)
Susmita Shrivastava

And the many other TIGR employees, present and past, who have
contributed to the development of these tools and to the annotation
protocols I have described.

Also thanks to the funding agencies that make our work possible
including NIH, NSF, and DOE.