This document covers building a TCW database for a single species. Terminology:
Note: there is still some old terminology floating around, such as the term "library" use to be used for both "dataset" and "condition",
and the acronym PAVE is still found in the configuration files as that was the name before TCW.
- Dataset is one of the following: (1) a set of sequences to assemble, with optional quality data
(2) a file of sequences with optional count data from conditions, where the sequences can be nucleotide or amino acid.
- Conditions may be tissues, treatments, etc that are to be compared, with optional replicates.
- AnnoDBs are fasta files of sequences (nucleotide or amino acid) to compare the dataset sequences
against for annotation. TCW provides special support for UniProt, but can use any file of sequences (e.g. Genbank nr).
|demoTra||Transcripts with counts, locations, remarks||Here - this section
|demoAsm||Assemble transcripts and ESTs||Assembly Guide
|demoPro||Protein sequences with counts||Same steps as for demoTra
Follow steps of Installation. Make sure your HOSTS.cfg is correct.
At the command line, type
The window shown on the right will be launched. Follow the instructions below on the left of the image.
Note that this demo project is pre-configured; to create a new project from scratch,
refer to Creating a New Project.
Select demoTra from the Project drop-down list.
Note that you can set the number of CPUs to use, although the
demo does not need more than one.
1. Click the Exec Load Data command
This loads the datasets in this section, where there is one
dataset of transcripts ("Ginger") and three
conditions with count data ("Root","Tip","Zone") where the first condition has five replicates
and the second two conditions have one replicate.
Note: A command button will turn gray while it is
executing, and there will be output to the terminal.
Keep an eye on the terminal because
you may be prompted with yes/no prompts which need to be answered.
If you take too long to respond, it times out; in which case, ctrl-C and restart.
Note: A command will not be active (i.e. grayed out) when it is not valid to run, e.g.
the Exec Instantiate command is grayed out since the datasets have not been loaded.
2. Click the Exec Instantiate command.
The Skip Assembly was checked so the transcripts will simply be loaded, without
The Use Sequence Names From File is not checked, meaning that the TCW will assign new,
sequentially-numbered names prefixed by the singleTCW ID.
From this point on, you may run
after any step to view the results that have been entered.
(Click to see larger image)
3. Click Exec Annotate Sequences
This searches against several UniProt partial databases which have been provided as part
of the package.
If the GO database has not been built yet (see Step 5), the following will be written to the terminal:
+++Warning: GO_tree go_demo is missing; ignoring GO step
--Please confirm above parameters. Continue with annotation? (y/n)?
Answer 'y' to continue.
The output to the terminal and to the file projects/demoTra/logs/annotator.log will look something like
this log (this includes the GO annotation).
4. Click the Add Remarks or Location button (bottom of window), a window will popup (not shown).
Select the file "traRemarks.txt" (when you press the "...", you will see the file),
then select Add Remarks from File.
This illustrates how to
add remarks to one or more sequences, which can be searched on in the TCW Basic search.
You may also add the "traLocations.txt" file to view locations in the resulting database.
5. To build the GO database, see
Demo annotation setup. Then execute Exec GO Only.
6. To compute differential expression (DE), install R and the respective packages. From the command line, execute
The DE Guide describes how to install the necessary
packages, and how to add DE p-values to the TCW database. If Step 5 has been run, then you can also add
the p-values for the GO.
Follow steps of Installation. Make sure your HOSTS.cfg is correct.
|To create a new project, press the Add Project button at the top of the |
You will be prompted for a project name; enter a name using letters, numbers, or underscores (no spaces).
Do not include "sTCW", that will be added. When you select "OK",
For example, if you enter the name "example", the ID will be "example" and the database will be "sTCW_example".
You cannot change the Database name, but you can shorten the singleTCW ID.
- The name will be entered for singleTCW ID, and for Database with the prefix "sTCW" .
- The libraries/<name> and project/<name> directories will be created with files LIB.cfg and sTCW.cfg, respectively. These maintain all information you enter.
singleTCW ID: This can be used as a command line parameter to
Also, if TCW generates
the sequence names (if Use Sequence Names from File is not checked), the singleTCW ID followed by sequential
numbers will be used.
Instead of using the Add Projects, you may create a directory under /libraries and put your sequence files along
with any other optional files (i.e. quality and count); when you select the project pulldown, you
will see your project (the project pulldown lists all directories under /libraries).
Define all datasets, then select Exec Load Data to load the data into the database.
Datasets cannot be added to an existing database.
Defining a sequence dataset
A TCW project must have at least one sequence dataset, which is a FASTA file of sequences.
Select Add beside the Sequence Datasets and the panel will be replaced with the one
shown on the lower right.
After entering the files and attributes, select Keep.
The main panel will reappear with the SeqID and Title added to
the Sequence Datasets table. Additionally, the information will be written in
SeqID: Enter the dataset name (a brief identifier) in the first box; if you want
to generate sequence names, it will use this identifier followed by consecutive numbers. It is
very important that you create very short descriptive name.
Add a more descriptive title and information
in the ATTRIBUTES section, which is shown on the Overview page of
Sequence File: Click the browse button labeled "..." to select the
fasta file of sequences.
The ">" lines are the sequence names, where characters other than letters, numbers, and underscores will be
changed to underscores.
Additional files may be added:
- For already assembled transcripts (e.g. Illumina) or protein sequences, there maybe associated count files.
Adding them is covered in the next section.
- For Sanger ESTs or 454 data, there may be quality data. Enter the name of the quality file.
- For Sanger ESTs, TCW assumes the 5' ESTs have the ".f" suffix and the 3' ESTs have the ".r" suffix. If there
are different from this, enter the correct ones.
Defining attributes and updating attributes
Enter any additional information
as desired in this section. This information will be shown on the Overview panel of
The attribute information can be added or changed after the database is created by using the Edit button on the main panel.
Note, they can only be changed from the directory which has the libraries/<name>/LIB.cfg as a subdirectory.
Defining count data
Associated with each dataset may be one or more conditions with count data.
Each condition has a count for each of the sequences.
Count File: Click the browse button labeled "..." to select the file containing the table of counts
for the sequences (see sample below).
If you have the replicate library counts in separate files, you can use
the Build combined count file option to create the 'table of counts' file, as discussed below.
Keep: The dataset panel disappears and the main panel returns. Note that the
sequence dataset is now shown in the first table, while the conditions (column headings from the count file)
are listed in the second table.
Use Define Replicates to group replicates and Edit Attributes to add information for the
conditions, which will be shown on the
Build combined count file
The Count File that is input to the Load Data routine is a tabular file where the first line contains
the word "SeqID" (or any label) followed by the condition column headings. The file must be space delimited; if your
file uses commas, they can be replaced with sed -i"" 's/,/" "/g' filename
Example Count File: This is part of the demoTra count file, showing the first 4 transcripts and
counts for three conditions, where the first has five replicates. The condition names and replicate numbers are automatically
determined from the column headers. Note that the sequences in sTCW_demoTra are prefixed with "tra_" and numbered
sequentially; this is because the 'Use Sequence Names from File' was not selected.
SeqID Root1 Root2 Root3 Root4 Root5 Tip1 Zone1
tZoR_000023 378 1002 1649 826 1195 726 151
tZoR_000117 101 206 151 109 185 129 58
tZoR_000246 1859 2506 1334 1541 2307 3012 976
tZoR_000335 529 919 1103 810 1427 2338 1438
However, it is common to have separate count files for each sample.
In this case, you can use the Build combined count file to generate the combined
To see how this works, create a new project and add a dataset
using the demoTra sequences, as shown above, but instead of adding the demo
demoTra.cnt.txt, select Build combined count file.
This brings up the interface shown at right (which shows the files already added).
All the files have been put in a sub-directory called "count":
Root1.cnt Root2.cnt Root3.cnt Root4.cnt
Root5.cnt Tip1.cnt Zone1.cnt
Note that each file name start with the condition name followed by the replicate
number. Since the files are all in one directory, and the
filename up to the first "." is the "library name + replicate number", we can use the bulk load.
Select Add Directory of Files, which brings up a file chooser window; select the directory containing
the files (e.g. "count")
and it loads all files from the selected directory, and as shown on the right.
If your files are not in one directory, or not named correctly, you will need to add them individually
using the Add Rep, as shown on the right. The Rep Name must be the (abbreviated) condition name followed
by the replicate number.
After adding all files, click Generate File, and the panel closes.
The Add/Edit panel returns, and now the Count File is filled in
with a file named "Combined_read_counts.csv".
As already described, the replicates are grouped by clicking the Define Replicates button on
the main window.
If the input is already assembled transcripts, or protein sequences, or gene sequences,
check Skip Assembly on the project interface.
Instantiation (with optional assembly)
|Go to top|
If you want the original sequence names to be used by TCW,
check Use Sequence Names from File; otherwise, TCW will rename the sequences
using a simple naming scheme. Note that, if the original sequence names are used,
characters other than letters, numbers, or underscores
will be replaced by underscores. Note also that the names supplied by
sequencers are often longer than necessary, in which case, it is better to
have the TCW assign names, which will use the library name followed by
consecutive numbers, hence, retaining their order information.
If the datasets are ESTs, or multiple transcript datasets that you want to assemble together,
do not select Skip Assembly. You can tune the assembly with Options, but it typically is not
In either case, you must press Exec Instantiate to either assemble or finalize the sequences in the database.
Annotation should be defined through the Manager, as in step 3 above.
Three types of computation are performed:
- Basic annotation: GC content and ORFs
- Functional annotation: Compares one or more protein and/or nucleotide databases ("annoDBs") to each sequence in the database.
- Similar sequences: Compute similar sequences, which is particularly useful for analyzing the results of
AnnoDBs and UniProt
The term "AnnoDB" refers to a fasta file of nucleotide or protein sequences,
where the TCW sequences (transcript or protein) will be blasted against the annoDBs for functional annotation.
Read Annotation Setup for obtaining UniProt and other annotation databases.
Import AnnoDBs provides a way to enter all UniProt databases at once.
If you have used
for the annotation setup,
make sure the last step was TCW.anno, which creates a file of UniProt databases.
Select Import AnnoDBs, which pops up a
file chooser, then select the projects/TCW.anno.<date>; all your UniProts will be added at once
with appropriate taxonomy values. Any additional databases (e.g. Genbank nr) need to
be added one by one.
Generate Hit Tabular File:
To add an annotation database, press the Add button next to the AnnoDB table. This brings
up the panel shown on the right.
Taxonomy: this does not need to be unique, but the DB type must be, which is created as follows:
DB type is
shown in various tables of
- Type: defined in the ">" of the annoDB fasta file of sequences, e.g. "sp" for SwissProt.
- Taxonomy: first 3 letters of the taxonomy, e.g. "pla" for plants.
- DB type: the type+taxonomy, e.g. if the type is "sp" and the taxonomy is "plant",
the TCW "DB type" will be "SPpla".
viewSingleTCW to indicate the origin of the hit.
Search Program: The drop-down will list TCW Select if
you have listed additional search programs (e.g.
diamond) in the HOSTS.cfg file.
The drop-down will also contain blast and any other search programs.
You can supply your own blast results file (must be in tabular format).
You must still provide the name of the annoDB fasta file as
it extracts the description and species from it.
- If the drop-down is set at TCW Select, then TCW will select which search program to use,
as described in selecting a search program.
- If you select a program, then you can also alter the parameters.
- If only blast is shown, then you have not identified any additional search programs in HOSTS.cfg.
The Options button below the AnnoDBs table provides additional options for:
The second two options are not available for TCW database created from protein sequences.
- Define the GO database for GO, KEGG, Pfam, and EC annotation.
- ORF finder parameters.
- Options for self-comparison of transcript sequences ("similar sequences").
GO Database (GO, KEGG, EC, InterPro, Pfam annotations)
See Annotation Setup for creating the GO database.
Once the GO database is created, it can be selected, as shown on the right. Once selected, it will display
the available GO Slim categories in the drop-down below it; alternatively, a OBO formated file may be entered.
ORF finder options
See the ORF document.
The annotator can also (optionally) compare all sequences and determine the top N pairs,
where their alignments can be viewed in
This can be helpful for assessing the stringency of an assembly, by noting how similar sequences are.
Adding to annotation
Additional annotation can be added at a later time; see Update and Redo annotation
in the Annotation Guide.
Adding remarks and locations
|...||Select to enter a file name.
|Add Locations from File||Reads the file, removes any existing location information,
and adds the location information to each sequence.
|Add Remark from File||Reads the file and adds the remarks to any existing remark
for the sequence.
|Remove Remark||Removes all remarks.
The file is a set of rows where the first word of each row is the sequence ID and the rest of the row is the format:
group:start-end(strand), e.g. SC_1:392-496(-)
- The group would be the supercontig, scaffold, chromosome, linkage, etc.
If TCW can extract numbers/X/Y from the end of the group (e.g. chr1, chrX),
it adds a column containing just the "Group" number that can be sorted numerically in
- The sequence ID must match a sequence ID in the database,
so you probably will want to "Instantiate" the sequences using "Use Sequence Names from File".
A script is available, scripts/extractCodingLoc.pl, to generate the transcript sequence file from a
genome sequence and GFF3 file. It also generates the location file in the format needed by TCW. It only works with
a subset of the GFF3 files, so probably needs to be modified
for your GFF3 or GTF file. It uses BioPerl.
The group name, start, end and strand are columns in the TCW database that can be viewed in
by selecting the "Columns" tab on the left, then checking the columns under "General".
The file is a set of rows where the first word of each row is the sequence ID and the rest of the row is the Remark.
- Single and double quotes will be changed to spaces.
- Semi-colon will be changed to a colon.
- Do not use the following remarks, as they are added during annotation:
Multi-frame; ORF contains Hit; Hit contains ORF; ORF overlaps Hit; !LG
viewSingleTCW, the Remark can be viewed by selecting the "Columns" tab on the left, then checking the "Remark"
column under "General". The remark can also be search on in the "Basic Queries Sequence"; this is a great way to add
additional information about your sequences. If you make the remarks "keyword=value", you can then search on the keyword
to get a specific group of sequences.
If errors occur, a message is written to the terminal
and the Java stack-trace is written to the file sTCW.error.log. This can
be sent to us so we can help you trouble shoot the problem, that is, if the
message to the terminal is not sufficient to indicate how to fix the problem. NOTE:
errors are appended to this file, so if an error keeps re-occurring without getting
fixed, the file can get quite large.
If the information on the
runSingleTCW panel does not look right, remove
/libraries/<project>/LIB.cfg to start over. You can try fixing the problem
by editing this file, but its necessary to format it correctly.
If that does not work, email email@example.com and we will guide you on how to enter the information.
If you have many sequences (e.g. transcripts) in the database and/or many annoDBs,
this can take a lot of memory. Running Exec Annotate Sequences sets the memory to 4096, which may not be
enough; in this case, once
runSingleTCW is ready to annotate your database, exit and run from the command line:
You can increase the memory size in the
TCW integrates several different R packages for computing differential expression of transcripts, including
DESeq2; an R script containing the R commands to compute DE can also be supplied. GOseq is supported for
computing DE enrichment of GO categories. The DE computations are pairwise, i.e. conditions are compared two
at a time. Each comparison results in a
column added to the database, which may be viewed and queried. The DE modules are accessed either
./runDE on the command line, or the 'Launch DE' button at the bottom of the
Manager interface (Fig. 1).
For full details on the DE modules, see the Differential Expression Guide.
Users do not ordinarily need to look at the underlying directories and files used by TCW, however they are described
here since it may be helpful at times.
Directory structure and configuration files
|Go to top|
When a project is added with Add Project,
runSingleTCW creates one directory each under /libraries and
/projects with the user supplied name (referred to here as <project>).
The libraries/<project> directory has a
LIB.cfg file where
runSingleTCW saves all the library information and the
has a sTCW.cfg file with all the assembly and annotation information.
Though you can put your data files anywhere that
runSingleTCW can access, you may want to put your data
files in the libraries/<project> directory in order to keep everything in one place.
Also, the /DBfasta directory contains a subset of the UniProt files for the demo; this is
a good location to put all annoDB files.
execAnno write to the projects/<project> directory, where
execAssm writes mainly log files, while
execAnno puts all
blast results in uniblast subdirectory. Blast files are NOT removed as they may be reused.
Both programs write log files to a log subdirectory.
As mentioned above,
runSingleTCW creates /libraries/<project>/LIB.cfg
and projects/<project>/sTCW.cfg. Once created, you can run the three steps from the
command line instead of through the interface.
The LIB.cfg and sTCW.cfg files can be edited with a text editor, but be careful as
runSingleTCW and the executables expects the syntax to be exactly as it writes the file.
Once a project is created, it is viewed and queried using
This program is launched either:
The TCW Tour shows the various displays.
- From the command line (./viewSingleTCW), which
brings up a panel of mySQL databases with the sTCW_ prefix, where databases can be selected to view.
- From the command line using the singleTCW ID or database name as a parameter (e.g. viewSingleTCW tra).
- Through the button at the bottom of the Manager interface (Fig. 1).
The following can be used to display
viewSingleTCW on the web with a given TCW database:
Running viewSingleTCW (please wait)
<param name="ASSEMBLY_DB" value="sTCW_your_db_name">
<param name="DB_URL" value="www.your.hostname">
<param name="DB_USER" value="your username (read-only)">
<param name="DB_PASS" value="your password">
<param name="ASSEMBLY_ID" value="your singleTCW ID">
<param name="DESCRIPTION" value="">
Unable to display the viewSingleTCW applet. Please verify your Java installation.
- Before using it on the web, run it as a desktop application as it may need to
update the Overview page.
That is, every time the database is changed, the Overview needs
to be updated, which is done when you execute
viewSingleTCW or the Overview option from
any of the "run" executables. If it is not updated before running from the web, the web applet will
not work because it cannot write to the database for overview update.
- Also, you may need to use a port other than the default
MySQL port; see MySQL port access.
Also, users may need to increase their allowed memory for Java applets,
see Out of memory.