BIO 331 - Structural and Functional Annotation

Genome annotation

Attaches biological information to sequences

Structural annotation

Process of identifying coding genes (an intron-exon structures) and non-coding genes (e.g. tRNAs)

Functional annotation

Attaches metadata structural annotations (e.g. which product is encoded in the gene)

This image describes the workflow for ___

Genome annotation

Programs for finding repeats and masking/annotating them

RepeatMasker, emboss NUCLEIC REPEATS

This output is from ___

RepeatMasker

This program is ___

RepeatMasker

This program is ___

emboss NUCLEIC REPEATS

This output is from ___

emboss NUCLEIC REPEATS

Structural annotation helps identify these genomic elements

Non-coding genes, regulatory motifs/promoters, coding genes

ncRNAs

Non-coding RNAs; RNA molecule that is not translated into a protein

tRNAscan-SE

Finds tRNAs

RNAmmer

Finds rRNA

Rfam

Finds tRNA, rRNA, mtRNA, snRNA, miRNA, etc.

This output is from ___

Rfam

What does ab initio mean?

From the beginning

Ab initio signals

Specific sequences that indicate the presence of a gene nearby (e.g. promoter)

Ab initio content

Properties of protein-coding sequence itself (e.g. start & stop

codons)

Tools for ab initio discovery

GLIMMER, GeneMarkS, Prodigal, ORF finder

What site is this?

GLIMMER

This output is from ___

GLIMMER

What site is this?

GeneMarkS

This output is from ___

GeneMarkS

Why do split reads exist in eukaryotic RNASeq data?

Intron splicing

Average vertebrate gene is ___ kb long

30

Average vertebrate coding sequence is approx. ___ kb long

1

Average vertebrate coding region consists of ___ exons of about ___ bp each

6, 200

Exon boundaries can be defined through ___

RNASeq

FOr highest accuracy, one should use ___ data rather than screening public databases

Experimental

Databases for determining gene function

UniProt, RefSeq, Pfam/TIGERFAM, user-provided set of annotated proteins

What website is this?

UniProt

This output is from ___

UniProt

True or false, UniProt provides hits to specific proteins?

True

Pfam is used to identify the ___ your protein belongs to

Family

Superfamily vs subfamily

Superfamily: large group of distantly related proteins

Subfamily: small group of closely related proteins

Building Pfam families

Seed alignment used to build a profile HMM

Profile HMM is searched against sequence databases

All matches scoring equal to or grater than a given threshold are considered as true members of the protein family

These members are added to the seed alignment to generate the full alignment and from there a consensus HMM and a consensus sequence

HMMs

Hidden Markov Models; used in many BI applications (gene/protein prediction, phylogenetic analysis, alignments)

HMMs are systems that move from state to state with ___ probabilities. Each state produces a new possible outcome.

Finite

Issue with HMMs

States that are responsible for a possible outcome are unobservable; only final outcome is observable

Two probabilities associated with HMMs

Transition and emissions

Transition probability

Probability of transitioning from one state (e.g. AA) to another

Emissions probability

Probability that a given AA exists at that position in the alignment

Comparing highly divergent protein sequences is best achieved through ___

HMMs

BLAST is based on ___ comparisons

Pairwise

Seq2EC

Takes an unknown protein and can predict the EC that would match the protein

BRENDA

Takes EC number and gives names of the protein that is usually assigned

This website is ___

StructRNA finder

This output is from ___

StructRNA finder

This website is ___

InterPro scan

This output is from ___

InterPro scan

This website is ___

BlastKOALA

BlastKOALA provides information about ___

KEGG orthologs (KOs)

InterPro scan provides information about ___

Gene ontology (GO)

This output is from ___

BlastKOALA