In April 2018 we are releasing Annocript2.0. It has a renewed implementation and few important changes! To see all of them go to the HISTORY page. In Annocript2.0 we experimentally produced an installation script which allows to install all dependencies automatically!
Annocript is a pipeline for the annotation of de-novo generated transcriptomes. It executes blast analysis with UniProt, NCBI Conserved Domain Database and Nucleotide division adding also annotations from Gene Ontology, the Enzyme Commission and UniPathways. Annocript also gives information about the longest ORF and the non-coding potential using external software. Annocript is also capable to identify putative long non-coding RNAs by using an heuristic based on homology and sequence features.
The inpatient may go directly at the INSTALL page. Then, you should visit the tutorial at TUTORIAL to learn how to use Annocript.
Annocript is a highly configurable tool to annotate and analyze transcriptomes. After the transcriptome annotation you may want to run an expression analysis. From the Annocript 2.0 we released a perl script to seamlessly perform the pipeline for the differential expression analysis using the output from Annocript. The package is called DEA (Differential Expression analysis with Annocript). It currently uses edgeR (Robinson et al, 2010) starting from the raw counts table. You can find it at https://github.com/frankMusacchia/DEA.
In the annocript_utils section you can also find a useful Python script for intelligent search of specific words inside user-selectable columns of the Annocript output. Other useful plots and analyses are in work in progress and will be made available as soon as possible.
This is a very simple guide, it is big only because everything is explained step-by-step but we promise that, after you installed and ran once, you will do it very rapidly the following times and you will never stop using it! We strongly suggest to keep open this guide while you are installing and running Annocript for the first time.
Examples of whole transcriptome annotations made by Annocript can be downloaded from http://bit.ly/15vnALW.
Please cite: Musacchia F, Basu S, Petrosino G, Salvemini M, Sanges R. Annocript: a flexible pipeline for the annotation of transcriptomes which can also identify putative long non-coding RNA
To start using it you must follow instructions in the GUIDE folder. Below is a simple introduction of what you will find there and how to find solutions to possible problems.
- Annocript has some specific purposes. To figure out what they are, please read the INTRO file;
- Since Annocript is a pipeline, it uses some software that you have to download and install. Please follow the instructions on how to do it in the INSTALL file. There you will also learn how to install Annocript;
- Every software has its thorns! To understand how to run Annocript the first time please read the TUTORIAL file;
- When you will be an advanced user you will want to use Annocript in a faster and smarter way. ADVANCED_USAGE explains how to begin to do this. Here, an HINT and TIPS section contains a lot of useful information on how to perform advanced operations;
- FAQ contains Frequently Asked Questions (refer to the FAQ whenever you have errors that block Annocript);
- OUTPUT contains a list of all the output that Annocript gives, the organization of folders and of the configuration files; It also explains what are the field of the tabular output of Annocript.
- HISTORY is the history of changes to Annocript from it's creation to the current release.
Please enjoy this product and send us questions and suggestions!
To work, Annocript needs:
- MySQL database server
- You also need a MySQL client.
- Perl
- BioPerl
- Python
- R
- Blast+
- Portrait
- dna2pep
Check for the tested versions in the INSTALL page.
Any user will wonder how much time and space Annocript needs. We can say approximate times and sure space dimensions. Annocript execution time depends strongly from what modules the user is executing and of course from the speed and memory of your machine. Here we only show the most important processes that are always taking the biggest part of the overall computational time.
The database creation needs to be done only once but it depends on what proteins db you are creating (Uniref or Uniprot_kb). If you decide to use Uniprot_kb, Annocript will fill an annotation table with both the BLAST results from Swiss-Prot and Trembl while if you are using Uniref, both Swiss-Prot and Uniref result will be shown. The database comprise also information about domains, GO, enzymes and pathways. Furthermore, databases to use for BLAST will be downloaded also (see the INTRO file for more information about this step).
Consider as a suggestion that, if your machine is not very recent and powerful you can consider to install Uniref database (Uniref50 takes the lowest time and space). Otherwise Uniprot_kb is a good choice.
We took times for the creation of the database from 3 different machines with different cores and RAM memory:
- Machine with 24 core and 96 GBRAM
- Uniprot_kb_2014_03 - 7 hours
- Uniref_2014_04 - 3 hours
- Machine with 24 core and 24 GBRAM
- Uniref 2013_01 - 9 hours
- Uniprot_kb 2014_03 - 21 hours
- Machine with 8 core and 8 GBRAM
- Uniref 2014_05 - 38 hours
The speed of the overall process depends greatly from the time required for the download of the huge databases. Here the times were taken by using approximately 1MB/sec in download but depending from the speed of you connection the download process may take few hours more.
The space needed to build the MySQL database is approximately:
- UNIREF 2014_04: 52.61GB (50.1GB for files; 2.4 GB MySQL db)
- UNIPROTKB 2014_04: 92.6GB (85.3GB for files; 7.3 GB MySQL db)
Newer protein databases will slightly increase the necessary space according to their growth in term of proteins numbers. You can save some space by choosing, during the configuration, to remove the downloaded .zip files.
- zip files in Uniref: 8.5GB
- zip files in Uniprotkb: 14.5GB
Programs execution strongly depends from the transcriptome and the database we are using. The number of sequences but mostly their mean length. We took the times of the blastx execution (the most expensive process) against the Uniref90 database with our default parameters (which you will find in the config_user.txt) on a machine with 24 threads (2 Intel(R) Xeon(R) X5660 at 2.80GHz and 96GB RAM). This machine permitted the multi-threaded run of blast and rpsblast. We always used 20 core. The followings are the times when depending from mean length of sequences.
It needed about:
- 6 hrs for 11351 transcripts of mean length 951 nucleotides;
- 13 hrs for 30346 transcrips of mean length 1245 nucleotides;
- 20 hrs for 17776 transcripts of mean length 1478 nucleotides;
- 23 hrs for 30339 transcripts of mean length 1427 nucleotides;
- 46 hrs for 98174 transcripts of mean length 965 nucleotides;
- 72 hrs for 64388 transcripts of mean length 2125 nucleotides.
Depending by the powerful of you machine, the BLASTx and the rpsBLAST execution may take few hours more.
If you get some error during the installation or the running of Annocript please see the FAQ page of the user guide or ask on the google group: https://groups.google.com/forum/#!forum/annocript