PDA, "Pipeline
Diversity Analysis", is a collection of programs and modules
mainly written in Perl that automatically can:
search for
polymorphic sequences in a large database, and
estimate their genetic
diversity.
PDA has a
user-friendly, web-based interface where the user can select the
sequences to be analyzed and the parameters to be used. Sequences can
be retrieved from either Genbank (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Nucleotide;
see the NCBI's
Disclaimer and Copyright)
or the DPDB database (http://dpdb.uab.es)
as a list of accession numbers or a set of organisms and/or genes.
Low quality sequences coming from large-scale sequencing projects
(i.e. working draft), where most missing data is, will be
excluded from the analysis. Alternatively, sequences can be introduced manually in Fasta or
Genbank formats. All sequences will be grouped by organism and gene,
and groups will be aligned using the ClustalW algorithm. After,
different analyses will be performed.
2. User
defined parameters in the PDA web form:
Input:
First, you have to
indicate whether you want the sequences:
to be retrieved from a
database, or
they will be given by
yourself in Fasta or Genbank formats.
In the case you want
the sequences to be retrieved from a database, you must first choose
the source database from which they will be obtained. At this moment,
two options are allowed: Genbank or
DPDB. You must know that DPDB is a
database of nucleotide sequences from the Drosophila genus, so you can
use it in the case you are interested with sequences from this
taxonomic group. Then, you have to choose whether you will introduce a
list of organisms and/or genes, or a list
of accession numbers corresponding to
the source database. Every new item must be in a new line. For
example:
Otherwise, if you want
to introduce the sequences yourself, you only have to paste them in
the form or use the Navigator to find the appropriate file in your
computer. The program can read two formats: Fasta and Genbank. Follow
the instructions given below:
Fasta
format:
Each new sequence begins with a line >HEADER. The sequence follows on the
next lines until the next >HEADER line is found. Note that you cannot specify
any sequence annotation using this format. However, you can specify
the organism and gene names using the following syntax in the header:
>Organism|gene
Example:
Genbank
format:
You can retrieve sequences in Genbank
format from the Genbank database. Note that each record must end with
a new line including exclusively two bars as in the example:
LOCUS AF461290
384 bp DNA linear
INV 30-SEP-2003
DEFINITION Drosophila auraria cytochrome oxidase II (COII)
gene, partial cds;
mitochondrial
gene for mitochondrial product.
ACCESSION AF461290
VERSION AF461290.1 GI:20805483
KEYWORDS .
SOURCE mitochondrion Drosophila
auraria
ORGANISM Drosophila auraria
Eukaryota;
Metazoa; Arthropoda; Hexapoda; Insecta; Pterygota;
Neoptera;
Endopterygota; Diptera; Brachycera; Muscomorpha;
Ephydroidea;
Drosophilidae; Drosophila.
REFERENCE 1 (bases 1 to 384)
...
Main
parameters:
PDA will analyze the
performed alignments in terms of
Polymorphism, Synonymous and
non-synonymous substitutions (1),
Linkage disequilibrium (2) and
Codon bias (3) by default. However, you can
disable the last three analyses. Linkage disequilibrium is estimated
in non-overlapping sliding windows. You can choose the window
length, which is set to 50 segregating sites by default, although the
maximum value that will be used on the web is 100 consecutive
segregating sites (be aware that windows may not be of the same length
in nucleotides).
You can also choose
which Gene regions you want PDA to
analyze separately (4). Regions are that of Genbank annotations,
which is also used in the DPDB database. CDS and exon
are chosen by default.
The
Minimum number of sequences per category
is the minimum number of sequences you for alignment (5). Number 2 is set by default.
The
Minimum ClustalW Score for pairwise comparisons is the minimum percentage of homology between each pairwise sequences comparison in a final
alignment (6). The aim of this parameter is to separate fragmented or
incorrectly annotated sequences into different subgroups, but it is also very useful to
separate, in different alignments, sequences of a same organism that come from well separated populations.
Sequences are never used in more than one subgroup.
The
Minimum
sequences length in the analyses
makes the program to exclude those
sequences shorter than the number of nucleotides written(7).
A
Warning
message will appear in the
Alignments section if the proportion of excluded sites within an alignment (excluding
end gaps) is greater than the introduced percentage (8).
The results will be
given in both HTML pages and in a relational
database. This can be either in MySQL
or Microsoft Access formats (9).
Finally, you can
introduce your e-mail address if you wish to receive a notification as
soon as the results are available in our server (10). Note that this is not
a required option because a unique ID will be assigned to your
submitted job, and you will be able to retrieve the results using this
ID. You will have to go to "Request results by
ID" (at the top of the PDA Home Page) and introduce the unique
ID in the box.
ClustalW
parameters:
Finally, you can modify the default
Clustal parameters, although these have been optimized for
polymorphism analyses.
Fast
pairwise alignment:
K-tup: Can
be 1 or 2 for proteins; 1 to 4 for DNA. Increase this to increase speed; decrease to improve sensitivity.
Window length:
The
number of diagonals around each "top" diagonal
that are considered.Decrease for
speed; increase for greater
sensitivity.
Score type:
The
similarity scores may be expressed as raw scores
(number of identical residues minus a "gap penalty" for each gap)
or
as percentage scores.If the
sequences are of very different
lengths, percentage scores make more sense.
Topdiag: The
number of best diagonals in the
imaginary dot-matrix plot that are considered.Decrease (must be
greater than zero) to increase speed; increase to improve
sensitivity.
Pairgap: The
number of matching residues that must be found
in order to introduce a gap.This
should be larger than K-Tuple Size.This has little effect on speed
or sensitivity.
Multiple alignment:
Gap open:Reduce
this to encourage gaps of all sizes;
increase it to discourage them.Terminal
gaps are penalized same
as all others except for END GAPS not being selected.BEWARE of making this too small (approx 5 or so); if
the penalty is too small, the program may prefer to align each sequence opposite one long gap.
End gaps: Here you can select if you want the terminal gaps to be
penalized or not.
Gap extension:
Reduce
this to encourage longer gaps;
increase it to shorten them.Terminal
gaps are penalized same as
all others.BEWARE of making
this too small (approx 5 or so); if
the penalty is too small, the program may prefer to align each sequence opposite one long gap.
Gap distances:
Penalization for the distance between gaps.
3. Estimation
Optimization Method:
After the grouping
and alignment of sequences, a further step (optional) is taken
before estimating the polymorphism parameters. It is referred here
as the Estimation Optimization Method:
First, PDA groups the
sequences by length, so that sequences in the same group must not
differ in more than the 20% of their length.
It calculates the amount
of informative sites in each accumulative group of sequences (e.j.
group 1 (longest sequences), groups 1+ 2, groups 1 + 2 + 3, etc.).
PDA will use the set of
sequences with the largest number of informative sites (in some
cases discarding the shortest sequences).
In this example, the
first four sequences would be assigned to group 1, and the last four
sequences to group 2. The number of informative sites (without gaps)
using the four first sequences (group 1) is:
Therefore, we will have
more informative sites by using the four long sequences only and
discarding the short ones. PDA would show the alignment with all the sequences, but would use
the four long sequences only to calculate the polymorphism estimates
(n = 4
in the results).
To distinguish which
sequences were used in the analyses from those which were discarded, PDA uses a color code:
●
for sequences that were included in
the estimates, and
●
for sequences that were NOT
included in the estimates.
You can find this
information in the ALIGNMENTS page.
If you always want to use all
the sequences of the alignments to calculate the estimates, unselect
the appropriate box in the initial form:
Use the
ESTIMATE OPTIMIZATION METHOD
4. PDA output:
The output of the
program is stored in our server ofr a week. It includes a set of HTML pages with the
results of the alignments and analyses, as well as a database with the
same information but also all the sequences used and their annotations.
HTML
output:
The main HTML results
page is divided in different sections:
The
input parameters are described in the
first section (organism/gene, input database, analyses performed,
etc.)
The second section
allows you to download the databaseand see the
LOG files (where the program
saves possible errors that could occur
during the analyses)
Then, you can see all
the alignments performed. If you
click on "View complete information of
alignments", all the information for each individual alignment
will be displayed in four sections (each corresponding to some
columns):
Alignment information
Alignment ID
Organism
Gene
Type of region
Region
Alignment quality
(Special caution must be taken on the alignments having these
parameters a high value)
% excluded sites in the alignment
(due to gaps or ambiguous
positions). If the value is greater than 30% (or the user defined
value), a warning message will appear, and the value will be marked in
red.
Minimum/Maximum
sequences lengths (% difference)
Alignment
Clustal align (text
file)
Fasta align (text
file)
Jalview align
(graphic viewer / editor)
Clustal output
DND Tree file
(phylogeny)
Sequences used
Genbank / EMBL
accession number (and links to both databases)
Assigned sequence ID
(identification number used in alignments)
Location (corresponding
to the original sequence from Genbank/EMBL)
Averages of the main statistics for
Polymorphism,
Synonymous and non-synonymous substitutions,
Linkage disequilibrium and
Codon bias, by gene regions (including
all the studied organisms and genes). You can view the complete
results for each alignment by clicking on the appropriate
"View complete "Analysis" results" link.
Histogram Maker Tool
MySQL /
Microsoft Access database:
The database can be
downloaded as a compressed file from the main HTML results page, or
can be requested directly to our server if the format chosen is MySQL
(see a quick help). Its structure
in tables is represented by the following figure:
Description of all the parameters:
Polymorphism
G+C
G+C
percentage
n
Number of
sequences
m
Number of
nucleotides in the alignment (alignment length)
start
First
nucleotide in the analysis
end
Last
nucleotide in the analysis
excluded
Number of
excluded positions in the analysis (gap or ambiguous positions)
analyzed
Number of
analyzed positions
S
Number of
segregating sites (S)
Nei 1987
S / m
Number of
segregating sites per nucleotide
Theta (from S)
Theta θ
per site (estimated from S)
Eta
Minimum
number of mutations (η)
Tajima 1996
Eta / m
Minimum
number of mutations η per site
Theta (from Eta)
Theta θ
per site (estimated from η)
Theta
Theta θ
per DNA sequence (estimated from S)
Tajima 1993
V Theta (norec)
Variance
of θ per DNA sequence (estimated from S) - without recombination
dV Theta (norec)
Standard
deviation of θ per DNA sequence (estimated from S) - without
recombination
V Theta (rec)
Variance
of θ per DNA sequence (estimated from S) - free recombination
dV Theta (rec)
Standard
deviation of θ per DNA sequence (estimated from S) - free
recombination
Theta per site
Theta θ
per site (estimated from S)
Nei 1987
V Theta_site (norec)
Variance
of θ per site (estimated from S) - without recombination
dV Theta_site (norec)
Standard
deviation of θ per site (estimated from S) - without recombination
V Theta_site (rec)
Variance
of θ per site (estimated from S) - free recombination
dV Theta_site (rec)
Standard
deviation of θ per site (estimated from S) - free recombination
FSM Theta (from pi)
Theta θ
per site (estimated from π) - Under a Finite Sites Model
Tajima 1996
FSM Theta (from S)
Theta θ
per site (estimated from S) - Under a Finite Sites Model
FSM Theta (from eta)
Theta θ
per site (estimated from η) - Under a Finite Sites Model
k
Average
number of nucleotide differences
Tajima 1983
Vst k (norec)
Stochastic
variance of k - without recombination
Vs k (norec)
Sampling
variance of k - without recombination
V k (norec)
Total
variance of k - without recombination
Vst k (rec)
Stochastic
variance of k - free recombination
Vs k (rec)
Sampling
variance of k - free recombination
V k (rec)
Total
variance of k - free recombination
Pi
Average
number of nucleotide differences per site (π)
Nei 1987
Pi_JC
Average
number of nucleotide differences per site (π) - with Jukes&Cantor
correction
Nei 1987;
Jukes and Cantor 1969
V Pi_JC
Variance
of π (Jukes&Cantor)
Tajima
Tajima's D
Tajima 1989
Synonymous and non-synonymous changes
# Stop codons
Number of
STOP codons
# Codons
Total
number of codons
# Sites
Total
number of sites (nucleotides)
# SS
Number of
Synonymous sites
Pi (SS)
π in
synonymous sites
Pi_JC (SS)
π (with
Jukes&Cantor correction) in synonymous sites
V Pi_JC (SS)
Variance
of π (with Jukes&Cantor correction) in synonymous sites
# NS
Number of
Non-synonymous sites
Pi (NS)
π in
non-synonymous sites
Pi_JC (NS)
π (with
Jukes&Cantor correction) in non-synonymous sites
V Pi_JC (NS)
Variance
of π (with Jukes&Cantor correction) in non-synonymous sites
Pairwise comparisons (between pairs of sequences)
SynDif
Number of
Synonymous differences
Nei and Gojobori 1986
NSynDif
Number of
Non-synonymous differences
SynPos
Number of
Synonymous positions
NSynPos
Number of
Non-synonymous positions
Ks
Number of
synonymous polymorphisms per synonymous site
Ka
Number of
non-synonymous polymorphisms per non-synonymous site
Ks_JC
Number of
synonymous polymorphisms per synonymous site - with Jukes&Cantor
correction
Ka_JC
Number of
non-synonymous polymorphisms per non-synonymous site - with
Jukes&Cantor correction
V Ks_JC
Variance
of Ks (Jukes&Cantor)
V Ka_JC
Variance
of Ka (Jukes&Cantor)
Linkage disequilibrium
# Polym sites
Number of
polymorphic sites
# Pairwise comparisons
Number of
pairwise comparisons (between pairs of polymorphic sites)
ZnS
R2 average
over all pairwise comparisons
Kelly 1997
Pairwise comparisons (between pairs of polymorphic sites)
Dist
Distance
between both compared sites (in nucleotides, considering alignment
gaps)
D
D
Lewontin and Kojima 1960
D'
D'
Lewontin 1964
R
R
Hill and Robertson 1968
R2
R2
SChi2
χ2
test
Sig (SChi2)
Significance of the χ2
test
Fisher
Fisher
test
Sig (Fisher)
Significance of the Fisher test
Codon Bias
# Codons
Total
number of codons
# Sites
Total
number of sites (nucleotides)
# Stop Codons
Number of
STOP codons
# Codons (no !)
Number of
codons excluding STOP codons
# Codons (no !, W, M)
Number of
codons excluding STOP codons and those coding for a unique
aminoacid (W=Trp and M=Met) (nuclear universal genetic code)
ENC
Effective
Number of Codons
Wright 1990
CAI
Codon
Adaptation Index (the submitted sequences their selves are
taken as the reference set)
Sharp and Li 1987
SChi2
Scaled
χ2
Shields 1988
G+Cc
G+C
content in all coding positions
Wright 1990
G+C2
G+C
content in second positions of codons
G+C3
G+C
content in third positions of codons
RSCU
Relative
Synonymous Codon Usage
Sharp 1986
Alignments quality:
You should pay special
attention at the quality of alignments and revise them after each
analysis. PDA offers two special parameters that give a value from 0
(the best) to 100 (the worst) to each alignment according to:
The proportion excluded
sites within an alignment due to gaps
or ambiguous bases (end gaps will not be taken into account). Should
this number be
greater than 30% (or the user defined
value), it will be marked in
red and a warning
message will appear in a small window. If it reaches the 100%, no analyses
will be performed on the alignment.
The length of the
shortest and the longest sequences, and the percentage of
difference between both sequences
You can find these
parameters in the Alignments section of the HTML output,
or in the Index_Analysis table of the database.
5. Histogram
maker tool:
This tool allows you to create
personalized histograms for every parameter estimated. You can:
Restrict the information to specific gene regions
Restrict the information to a organism and/or gene
Choose the type of distribution (which parameter you want to
represent in the histogram)
Choose the order of the categories: histogram or frequency
Set a number of categories (number of bars in the histogram or
frequency representation)
6. Download
PDA:
The source code of PDA
can be downloaded from our FTP site under the
GNU General Public License (GPL) and be used locally. This is
highly recommended for big analyses. Please, go to the section
"Download Source Code" and register in
order to obtain a username and password and get access to the FTP
server. See the Installation file that comes with the distribution
or that can be downloaded here
(1,238 KB)for
specific instructions on how to use the program locally.
