Large-scale biology projects such as the sequencing of the human genome and gene expression surveys using RNA-seq, microarrays and other technologies have created a wealth of data for biologists. However, the challenge facing scientists is analyzing and even accessing these data to extract useful information pertaining to the system being studied. This course focuses on employing existing bioinformatic resources – mainly web-based programs and databases – to access the wealth of data to answer questions relevant to the average biologist, and is highly hands-on. Topics covered include multiple sequence alignments, phylogenetics, gene expression data analysis, and protein interaction networks, in two separate parts. The first part, Bioinformatic Methods I (this one), deals with databases, Blast, multiple sequence alignments, phylogenetics, selection analysis and metagenomics. This, the second part, Bioinformatic Methods II, covers motif searching, protein-protein interactions, structural bioinformatics, gene expression data analysis, and cis-element predictions. This pair of courses is useful to any student considering graduate school in the biological sciences, as well as students considering molecular medicine. Both provide an overview of the many different bioinformatic tools that are out there. These courses are based on one taught at the University of Toronto to upper-level undergraduates who have some understanding of basic molecular biology. If you're not familiar with this, something like https://learn.saylor.org/course/bio101 might be helpful. No programming is required for this course.
Lecture
Loading...
來自 University of Toronto 的課程
生物信息学方法 I
603 個評分
從本節課中
NCBI/Blast I
In this module we'll be exploring the amazing resources available at NCBI, the National Centre for Biotechnology Information, run by the National Library of Medicine in the USA. We'll also be doing a Blast search to find similar sequences in the enormous NR sequence database. We can use similar sequences to infer homology, which is the primary predictor of gene or protein function.
與講師見面
Nicholas James ProvartProfessor
Cell & Systems Biology
Okay. Welcome to Bioinformatic Methods I.
I am your instructor, Nicholas Provart.
The course material for this course was developed by Ryan Austin, David Guttman,
Laura Hug, Momoko Price and myself.
And just a reminder: please do use the Coursera tools to discuss lecture content
and labs, thank you.
So this course will cover the basics of searching one of the main
repositories of nucleotide and protein sequence information: NCBI -
the National Center for Biotechnology Information's Genbank, using GQuery or
Entrez, and Blast, which stands for Basic Local Alignment Search Tool.
We'll also create multiple sequence alignments and phylogenies, and
we'll cover selection analysis and "next gen" sequence analysis, and metagenomics.
Most of the tools we'll be using are web-based.
So here's an overview of the modules that we'll be undertaking.
And just as a note, there is a Bioinformatic Methods II, where we cover
protein motifs, protein-protein interactions, protein structure,
gene expression analysis, as well as cis regulatory elements.
The format of the course will be as follows.
It will consist of a 20 minute mini-lecture,
short two minute introduction and summary videos,
the weekly labs that are the focus of this course,
which will take between one and two hours to complete.
There'll be lab quizzes that go along with these, and
as well optional lab discussion videos.
So if you get stuck during the lab, you can turn to those videos for help.
There are two sectional quizzes, one after the first three weeks, so here.
And then one at the end of the course and
there's one final assignment which is due at the end of the course too.
So what is bioinformatics?
Bioinformatics is the development and
application of computational tools in managing all kinds of biological data.
It involves the use of computers for storage, retrieval,
manipulation, and distribution of information related to
biological molecules such as DNA, RNA, proteins, and metabolites.
It's generally limited to sequence, structural,
and functional analysis of genes and genomes and their corresponding products.
It's sometimes called computational molecular biology.
This field has really exploded over the past decade or so.
To help manage the huge increase in data generated by human and
other genome sequencing projects, high-throughput technologies, and so on.
So why do we need bioinformatics?
Well if you can imagine 3 billion letters in the human genome, 3
billion nucleotides, how do we really make sense of that without using computers?
So this is just a small section of a human genome,
encompassing the human globin gene.
And we would like to know which parts of the genome are important,
that code for proteins.
And without using computers, we would never know that this region here or
this region here comprised an exon of the globin gene.
(That is the piece of the gene that actually codes for protein.)
So we'll talk about biological databases a little bit here...
talk about why do we need databases, what is a database, data structures,
flat file versus relational databases.
We'll talk about accesion numbers and identifiers used in NCBI's Genbank and
other databases, and we'll go over a practical example of utility for
querying these databases using GQuery, sometimes called Entrez.
So why do we need databases?
When we look down at the Earth from several thousand kilometers away, we see
that it's covered with some green stuff that's actually indicative of plant life
and then there is also obviously an awful lot of animal life on the planet itself.
And the idea is with a universe of information that we're now able to
generate - actually the background here is an output of an Illumina
sequencing machine for generating sequence information -
we can now generate a universe of information and
understand different aspects of life better.
So, we might want to capture, for instance, in databases, genome and
genomic sequences, gene sequences and mutations thereof.
We might want to touch on gene regulation, how genes are regulated,
when and where genes are expressed, the transcript levels.
We might be interested in intron splice variants,
which could actually lead to disease.
We're certainly interested in protein sequences because proteins are the things in our body
that actually fold up into small molecular machines that do a lot
of the aspects that are required for life.
We might be interested in post-translational modifications or
protein tertiary structure.
Proteins often don't operate in isolation, rather they operate in networks.
And we need to know what the interactors are within those networks
to better understand the biology.
Where a protein is localized is important for its function.
Enzyme kinetics are important for a subset of proteins that are enzymes.
There also databases where we could look at metabolites that are altered
in response to different environmental perturbations or under different diseases.
Of course there are databases of disease and to put an academic framework around
everything, we would like to be able to search the literature.
So basically, we need databases to archive accumulated knowledge and
to provide scientists with easy access to biological data.
So what is a database, or how could we store data?
We could store data in a flat file format with
the fields separated by some delimiter.
So here is just a small database of four professors.
We've got first name, this pipe character, that's the delimiter,
last name, the department, University of Toronto, and then the address.
So that's one way of storing data, in a flat file format.
We could also store these data in a spreadsheet,
and you're familiar with Excel I'm sure.
We could have a column for the first name, column for the last name,
column for the institution, column for the department, and an address.
There are problems with this sort of database, however, and
one of these is redundancies.
So you'll here that we've actually stored the same piece of information twice and
this takes up storage space, and even though storage is getting cheaper and
cheaper still this means that when the physical location moves we have to update
all of these records, and
we have to ensure that we update all of those records.
Also, there is some issue here related to not storing
the department name in the same manner.
Something called relational databases offer a solution, and
in the case of a relational database, we have a series of relations or
tables which contain attributes, which are fields or columns.
And each row in a table is known as a tuple or a record, and
within this database the data should be normalized so that there isn't redundancy.
So every row should be unique in other words,
although this ideal isn't always observed.
So in the case of our professor example, what we can do is actually
separate out the table of professors from the table of the information
regarding where they reside or which department they're associated with.
So we could come up with a table of professors.
Here's our Professor_id, it's the primary key, we assign a number.
We have a First_name, a Last_name, and then we can link
this Professor table to the table of Contacts using a foreign
key which links to the primary key in the Contacts table.
And here what we do is we have Contact IDs 1,
2, 3, these are the primary key for this table.
And here we're indexing the Institution, Department, and the Address.
So these kind of relational databases can
be created with a language called structured query language, SQL.
And a lot of biological databases are built using SQL.
So, within a biological database, there may be a couple of different ways of
identifying a given entry, either with an identifier or an accession code.
And in the case of an identifier,
this corresponds to a locus in GenBank or entry name in UNIPROT,
this consists of a string of letters and
digits that are understandable in some meaningful way by a human being.
They aren't as stable as accession numbers
because they can be modified by curators if the presumed
function of the protein is found to be something else. For
instance in the case of ADH6 (alcohol dehydrogenase 6) gene in
human the identifier in UNIPROT looks like this, ADH6_human.
And in GenBank it looks like this SEG_HUMADH6A0.
So if we were to look at that say in a multiple sequence alignment
using those identifiers we could at least understand that the given sequence
say, it came from a human, and that it corresponded to ADH6.
So, that's kind of useful for labeling sequences and
multiple sequence alignments.
And as I mentioned, identifiers can change, and
the database creator might decide that the entry is no longer appropriate, and
it doesn't happen that often, but it does happen for our example.
So it has happened for our example.
It used to be this identifier, and now it's this identifier.
So the other way of identifying sequence is using an accession code or
a number, and that's a number with perhaps a few characters in front that
uniquely identifies an entry, and it's often assigned arbitrarily.
And in the case of ADH6_HUMAN in UNIPROT.
The accession code is P28332, so
it is not very intuitive as to what the actual sequence does
in relation to its function, and in the case of the GenBank,
the accession code for the ADH6 gene is AH001409.
Again, not at all intuitive as to what that sequence might do.
Another important thing to know about databases is that
sequences can be updated, and when they're updated, new identifiers are created.
In 1992,
the NCBI began assigning a unique number for each sequence that was submitted.
The GenInfo identifier or GI number, and this in the case even for
updated versions of a given accession, unless the two sequences were identical,
the same accession number may be associated with a different GI
if a newer or corrected sequence is submitted.
In the case of GenBank records, we typically have an Accession.Version
identifier, such as this, M84402.1.
".1" is the version.
And this identifier is mapped to its unique corresponding GI number,
which can be considered the primary key of GenBank.
So, the key point here is that to specify a sequence exactly in GenBank we
use either its GI number or the Accession.Version.
Conversely, to retrieve the most up to date sequence we
can use the accession number without the version.
And in that case, the most up-to-date sequence will be retrieved automatically.
So let's look at the GenBank record for human alcohol dehydrogenase VI.
Here's the link for it and this case is actually keyed by its GI number and
here's the NCBI record.
And if we look at the format that's returned to us,
this is the GenBank Flatfile Format, the GBFF, and
this is one of the most commonly used formats for nucleotide sequences.
It contains all of the information associated
with the sequence, as well as the sequence itself.
So it's got three parts to it.
The header, the features, and the sequence itself.
In the header we're looking at the locus which tells you the identifier,
the length of that sequence, the source, whether it's linear or
circular, the taxonomic group, here, primates
(this is actually deprecated), and the entry date,
which is the 17th of October, 2000.
In the header, we've got some further information and
a Definition, which is the biology of the molecule in a sentence.
We've got the accession codes, the version which is a number, the dot one.
And the GI number is found on this line too.
And in some cases we have keywords as defined by the submitters.
But the problem with this is that these are not at all structured.
There's no structured keyword system for
GenBank.
They're just defined by the submitters.
So just be aware of that when you're searching.
Also in the header we would have the Source which contains the organism name.
The Organism which contains complete taxonomic information from
the NCBI taxonomy server.
And sometimes we have references which contain details on
the publication about that sequence.
And maybe a comment
containing miscellaneous information and revision details.
The next part of the GenBank Flatfile Format is the features and the features
are a direct representation of the biological information in the record.
The Source Feature must be present in all GenBank records and should contain
information as to where the molecule comes from, so in this case Homo sapiens.
It might also contain, potentially, map, chromosome, and tissue type information.
The exon feature tells us that the sequence from
287 to 396 comprises an exon which is the first exon.
In some records, we have the CDS or the coding sequence feature and
this provides a translation of the sequence
after we've removed introns - we'll get to this in the lab -
a translation of the nucleotide sequence into its corresponding protein sequence.
And then the last part of the GenBank flat file record is the sequence itself,
starting up here with TGT and ending down here with TTT.
So GenBank has really grown dramatically over the years and we're pushing,
we're close to a trillion bases as of 2015.
GenBank actually doesn't contain every single
nucleotide sequence that has been submitted to the NCBI.
There are also a bunch of whole genome sequences that haven't yet
been incorporated into GenBank.
The number of those is even higher than the ones in GenBank.
And this is just the graph for the number of bases.
Here's the graph for the number of sequences.
We're approaching a billion sequences in GenBank, and
again the number of whole genome shotgun sequences has surpassed that number too.
So how can we search this enormous volume of information?
Well, we can use a keyword search.
So we'll be doing that in the lab.
We'll be exploring how to search using keywords.
We can also search by sequence similarity, using a tool called BLAST,
Basic Local Alignment Search Tool.
And just as a note here, Google and
other search engines don't really handle sequence searches well.
And this is because they can't put gaps in to identify partial matches to similar
sequences.
And they also don't know which amino acids have similar properties.
And we'll talk a little bit more about how BLAST works in the next lecture.
But this is why we have to use a specialized search tool for
searching sequenced databases.
Also before we start the lab, I should define some terms.
Let's consider an early globin gene and
this has occurred in an ancestral organism.
Start off with a single copy of this gene here and
then there's a gene duplication event and
this can happen through a variety of ways it's like segmental duplication, or
whole genome duplication.
Typically, in this case, what we would be looking at is segmental duplication
such that there's now an alpha version of the globin gene, and the beta version.
And then, at some point down the road, we've had
a speciation event and leading to frogs,
chicks, mice and other organisms and
in this case we would have a copy of the alpha globin in frog, chick, and mice.
A copy of the beta globin in mouse, chick, and frog and what do we call,
how are those sequences related to one another, what are the terms that we use?
So all of the alpha globins are orthologs of one another,
all of the beta globins are orthologs of one another.
The mouse alpha and the mouse beta are paralogs of one another.
And all of those sequences together are homologs, so
just keep that in mind when youre specifying sequences and
the relationships one to another... something to think about.
So, another way of searching across databases,
in addition to using Blast, is to use these overarching search tools.
So in the case of NCBI, there's a system called GQuery, or Entrez.
And what we'll be doing is, we'll be going through an example using this Entrez or
GQuery, a framework for querying the various databases that are NCBI.
Let's say we have a sample problem which is to identify SNPs,
single nucleotide polymorphisms, which potentially cause early onset
breast cancer and design oligonucleotides to PCR them,
the sequences containing them, in samples of human genomic DNA for sequencing.
So this is kind of a screening method.
So we can start with the OMIM function of Entrez or
GQuery and OMIM has links to everything that's
known about a given disease across the various databases at NCBI.
So you probably can't see this in the video, but
there are extensive cross linkages between the various databases at NCBI and
we can use this Gquery framework to get at,
to go from the, say, disease database to the SNP database quite easily.
So here we are at OMIM and we'll be working through an example in the lab.
In this case we start OMIM.
We type in early onset breast cancer and then one of the entries that comes up is
one sort of a well-studied gene that's involved, that's linked to breast cancer.
What we can do is just click on this gene summary link down here
and this will take us to a summary within
the gene part of NCBI for that gene, for BRCA1.
Then, already we can see in this part of the output of BRCA1,
on the output of BRCA1 gene page, some polymorphisms.
If we click here, on this link, the gene view, the SNP gene view link, we're
taken to a page which shows us all of the polymorphisms that have been identified,
both clinically relevant polymorphisms and from genome sequencing projects
(there are many more than a single genome available for humans now)
that result in changes that would result in
changes to the protein-coding sequence for that gene, and does, potentially,
disrupt the function of the BRCA1 protein, which is a protein for DNA repair.
So all of the SNPs in red, actually,
result in changes to the protein-coding sequence.
So what we would want to do is to amplify a piece of
DNA containing the region where those SNPs would occur.
And we can very easily go to the part of the NCBI that contains the DNA.
A contig, and there it is.
It's on chromosome 17.
We can specify region to retrieve the sequence information for
by clicking on the update view.
And we can retrieve that sequence in FASTA format,
which is a common format for sharing sequence information.
And there we go.
We've got our region of sequence.
And now all we have to do is design primers to amplify that region of DNA so
when we sequence someone's amplified genomic DNA, we could see whether or
not there are polymorphisms there for breast cancer susceptibility.
And we can use a tool called Primer3,
where we paste that sequence in FASTA format into the box,
we specify the size of the PCR product that we would want, and then we click go.
And there is our, there's a primer here that Primer3 has chosen for us.
And there's another primer here that we can't see on the screen.
Which is the right primer that we need for PCR amplification.
And this is just a new variation view
that NCBI is rolling out for looking at polymorphisms.
And the same sort of steps would apply here and we can quite easily go
to the genomic region and again retrieve the FASTA sequence.
So it's not quite rolled out but I'm just including it for sake of completeness.
So that's the end of our mini-lecture.
The important thing to think about when we're talking about databases
is which database do we use for what.
Over the course of this course,
we will actually be encountering different databases, some for
protein sequences, some for DNA sequences, some for protein-protein interactions
if you take Bioinformatics 2, and it really is a, there is an incredible
wealth of information out there if we only know which databases to access and how to
use the tools to access the information in the database correctly and specifically.
So that's what we will be doing in the lab today.
And I hope you enjoy it.
And we'll see you next time.
Thanks!
