MWAS Tutorial

Jump to: navigation, search

Maker Web Annotation Service

The MAKER Web Annotation Service (MWAS) is an easily configurable web-accesible genome annotation pipeline. It's purpose is to allow research groups with small to intermediate amounts of eukaryotic and prokaryotic genome sequence (i.e. BAC clones, small whole genomes, preliminary sequencing data, etc.) to independently annotate and analyse their data and produce output that can be loaded into a genome database. MWAS is build on the stand alone genome annotation pipeline MAKER, and users who wish to annotate datasets that are too large to submit to MWAS are free to download MAKER for use on their own systems.

Understanding MWAS

The first half of this page gives general background to genome annotation as well as describes validation data for the MAKER Web Annotation Service, MWAS. The stand alone annotation pipeline MAKER is at the heart of MWAS, and MWAS has been configured to present the user with configuration options that match those of the command line program MAKER as closely as possible.

Introduction to Genome Annotation

What Are Annotations?

Annotations are descriptions of different features of the genome, and they can be both structural or functional in nature.


  • Structural Annotations: exons, introns, UTRs, splice forms etc.
  • Functional Annotations: process a gene is involved in (metabolism), molecular function (hydrolase), location of expression (expressed in the mitochondria), etc.

It is especially important that all genome annotations include with themselves an evidence trail that describes in detail the evidence that was used to both suggest and support each annotation. This assists in quality control and downstream management of genome annotations.

Examples of evidence supporting a structural annotation:

  • Ab initio gene predictions
  • ESTs
  • Protein homology

Importance of Genome Annotations

Why should the average biologist care about genome annotations? Genome sequence itself is not very useful. The main question when any genome is sequenced is, "where are the genes?" To identify the genes we need to annotate the genome. And while most researchers probably don't give annotations a lot of thought, they use them everyday.

Examples of Annotation Databases:

Every time we use techniques such as RNAi, PCR, gene expression arrays, targeted gene knockout, or CHIP we are basing our experiments on the information derived from a digitally stored genome annotation. If the annotation is correct, then these experiments should succeed; however, if an annotation is incorrect these experiments are bound to fail. Which brings up a major point:

  • Incorrect and incomplete genome annotations poison every experiment that uses them.

Quality control and evidence management are therefore essential components to any annotation process.

Effect of Next Generation Sequencing on the Annotation Process

It’s generally accepted that within the next few years it will be possible to sequence even human sized genomes for as little as $1,000 and in a short time frame. When these expectations finally become reality, then whole genome sequencing will likely become routine for even small laboratories. Unfortunately, advances in annotation technology have not kept pace with genome sequencing, and annotation is rapidly becoming a major bottleneck affecting modern genomics research.

For example:

  • As of October 2009, 222 eukaryotic genomes were fully sequenced yet unpublished (this is an ever growing backlog).
  • Currently (Jan 2010) there are over 900 eukaryotic genome projects underway, assuming 10,000 genes per genome, that’s 9,000,000 new annotations (with this many new annotations, quality control and maintenance become an issue).
  • While there are organizations dedicated to producing and distributing genome annotations (i.e ENSEMBL and VectorBase), the shear volume of newly sequenced genomes exceeds both their capacity and stated purview.
  • Many small research groups (which often lack bioinformatics experience) must therefore confront the difficulties associated with genome annotation on their own.

The MAKER Web Annotation Service is a tool to assist research groups in converting the mountain of genomic data provided by next generation sequencing technologies into a usable resource, and for larger datasets, research groups can use a local installation of the annotation pipeline MAKER.

What does MWAS do?

  • Identifies and masks out repeat elements
  • Aligns ESTs to the genome
  • Aligns proteins to the genome
  • Produces ab initio gene predictions
  • Synthesizes these data into final annotations
  • Produces evidence-based quality values for downstream annotation management

File:Apollo view.jpg
MAKER generated annotations, shown in Apollo.

What sets MAKER and MWAS apart from other tools (ab initio gene predictors etc.)?

MAKER is an annotation pipeline, not a gene predictor. MAKER does not predict genes, rather MAKER leverages existing software tools (some of which are gene predictors) and integrates their output to produce what MAKER believes to be the best possible gene model for a given location based on evidence alignments.

gene prediction ≠ gene annotation

  • gene predictions are gene models.
  • gene annotations are gene models but should include a documented evidence trail supporting the model in addition to quality control metrics.

This may seem like just a matter of semantics since the primary output for both ab initio gene predictors and the MAKER pipeline is the same, a collection of gene models. However there are a few very significant consequences to the differences between these programs that I will explain shortly.

Emerging vs. Model Genomes

Emerging model organism genomes each come with there own set of issues that are not necessarily found in classic model genomes. These include difficulties associated with Repeat identification, gene finder training, and other complex analyses. Unfortunately emerging model organisms are often studied by very small research communities which often lack the resources and bioinformatics experience necessary to tackle these issues.

Classic Model Organisms Emerging Model Organisms

Well developed experimental systems

New experimental systems

  • Genome will be the central resource for work in these systems

Much prior knowledge about genome

Little prior knowledge about genome

  • Usually no genetics
Large community Small communities
Big $ Less $
Examples: D. melanogaster, C. elegans, human, etc. Examples: oomycetes, flat worms, cone snail, etc.

Comparison of Algorithm Performance on Model vs. Emerging Genomes

If you have ever looked at comparisons of gene predictor performance on classic model organisms such as C. elegans you would conclude that ab initio gene predictors match or even outperform state of the art annotation pipelines, and the truth is that, with enough training data, they do. However, it is important to keep in mind that ab initio gene predictors have been specifically optimized to perform well on model organisms such as Drosophila and C. elegans, organisms for which we have large amount of pre-existing data to both train and tweak the prediction parameters.

Table: MAKER's Performance on the C. elegans genome



Ab initio Evidence Based
SNAP Augustus MAKER Gramene
Genomic Overlap (gene)
SP 82.48 88.09 91.69 93.49
SN 95.44 96.78 89.81 88.74
Exon Overlap
SP 18.88 22.87 25.58 27.38
SN 87.63 93.09 91.17 94.84

What about emerging model organisms for which little data is available? Gene prediction in classic model organisms is relatively simple because there are already a large number of experimentally determined and verified gene models, but with emerging model organisms, we are lucky to have a handful of gene models to train with. As a result ab initio gene predictors generally perform very poorly on emerging genomes.

MAKER's Performance on the S. mediterranea Emerging Model Organism Genome. Pfam domain content of gene models determined using rpsblast

By using ab inito gene predictors inside of the MAKER pipeline instead of as stand alone applications you get certain benefit:

  • Provide gene models as well as an evidence trail correlations for quality control and manual curation
  • Provide a mechanism to train and retrain ab initio gene predictors for even better performance.
  • Output can be easily loaded into a GMOD compatible database for annotation distribution (including evidence associations).
  • Annotations can be automatically updated with new evidence by simply passing existing annotation sets back into the pipeline

Getting Started with MWAS


MWAS is free to all users for academic use and has no login requirement, but registration is recommended as it allows for easier file and job management and registered users are allowed to upload more sequence.

Running MWAS with Example Data

MWAS comes with some example files to familiarize the user with how to run an annotation job. You can pre-load the fields for an example job by selecting one of the examples from the drop down menu on the "New Job" page and then selecting "Load". This will fill out options on the "New Job" form for you. Review the options carefully, and then submit the example job for execution by pressing the "Submit to Queue" button at the bottom of the page.

Start with the "Drosophila melanogaster : DPP example". This will load the region of the D. melanogaster genome encoding decapentaplegic along with cDNA and protein evidence overlapping the region. Select "Drosophila melanogaster : DPP example" from the drop down example menu. Then select load to fill in the form.

If you scroll down through the form, you will notice that the genome file, EST file, protein file, and prediction method sections have been filled out for you. Click on "Submit to Queue", to start the job.

You should be redirected to the MWAS start page upon submisssion, and the job you have submitted should be visible in the job status section. Click "Refresh Job Status" to update the run status of your job. Within a few moments, your job will complete, at which point you can view the results

Click on "View Results". You can now download the results for local analysis on your own system or you can click on "View in Apollo" to seen gene models loaded directly in the Apollo genome browser. This option will install a Java Web Start version of Apollo if it is not already installed. You can also view summery statistics of the annotation from the Sequence Ontologies SOBA tool by clickin on "SOBA Statistics".

Details of What is Going on Inside of MWAS

Repeat Masking

MWAS runs MAKER internally, an the first step to MAKER is repeat masking, but why do we need to do this? Repetitive elements can make up a significant portion of the genome. Some of these repeats are simple/low-complexity repeats where you have runs of C's or G's or maybe even something like AAGGAAGGAAGG. Other repeats are more complex, i.e. transposable elements. These high-complexity repeats often encode real proteins like rerotranscriptase or even Gag, Pol, and Env viral proteins. Because they encode real proteins, they can play havoc with ab initio gene predictors. For example, a transposable element that occurs next to or even within the intron of a real protein encoding gene might cause a gene predictor to include extra exons as part of a gene model, sequence which really only belongs to the transposable element and not to the coding sequence of the gene. You will also get hundreds of instances where identical transportable element proteins get annotated as being part of an organisms proteome. In addition these issues, low-complexity repeat regions can align with high statistical significance to low-complexity protein regions creating a false sense of homology throughout the genome. To avoid these complications it is convenient to identify and mask any repeat elements before doing other analyses.

MAKER identifies repeats in two steps.

  • First a program called RepeatMasker is used to identify low-complexity and high-complexity repeats that match entries in the RepBase repeat library, or any species specific repeat library supplied by the user.
  • Next MAKER uses RepeatRunner to identify transposable element and viral proteins from the RepeatRunner protein database. Because protein sequence diverges at a slower rate than nucleotide sequence, this step helps pick up the most problematic regions of divergent repeats that are missed by RepeatMasker, which searches in nucleotide space.

Regions identified during repeat analysis are masked out so as not to complicate other downstream annotation analyses.

  • High-complexity repeats are hard-masked, a technique in which nucleotide sequence is replaced with the letter N to prohibit any alignments to that region.
  • Low-complexity regions are soft-masked, a technique in which nucleotides are made lower case so they can be treated as masked under certain situations without losing sequence information. I will discuss some of the applications and effects of soft-masking later.

Now the idea of masking out sequence might seem on the surface like we're losing a lot of information, and it is true that there can be proteins that have integrated repeats into their structure, so repeat masking will affect our ability to annotate these proteins. However, these proteins are rare and the number of gene models and homology alignments improved by this step far exceed the few gene models that may be negatively affected.

Ab Initio Gene Prediction

Following repeat masking, MAKER runs ab initio gene predictors specified by the user to produce preliminary gene models. Ab initio gene predictors produce gene predictions based on underlying mathematical models describing patterns of intron/exon structure and consensus start signals. Gene models are not produced by directly using experimental evidence. Because the patterns of gene structure are going to differ from organism to organism, you must train gene predictors before you can use them. I will discuss how to do this later on.

MAKER currently supports:

  • SNAP
  • Augustus
  • GeneMark
  • FGENESH (Disabled on public MWAS site)

You must specify HMM files you want to use use when running each of these algorithms.

EST and Protein Evidence Alignment

A simple way to indicate if a sequence region is likely associated with a gene is to identify (A) if the region is actively being transcribed or (B) if the region has homology to a known protein. This can be done by aligning Expressed Sequence Tags (ESTs) and proteins to the genome using alignment algorithms.

  • ESTs are sequences derived from a cDNA library. Because of the difficulties associated with working with mRNA and depending on how the cDNA library was prepared, EST databases usually represent bits and pieces of transcribed mRNAs with only a few full length transcripts. MAKER aligns these sequences to the genome using BLASTN. If ESTs from the organism being annotated are unavailable or sparse, you can use ESTs from a closely related organism. However, ESTs from closely related organisms are unlikely to align using BLASTN since nucleotide sequences can diverge quite rapidly. For these ESTs, MAKER uses TBLASTX to align them in protein space.
  • Protein sequence generally diverges quite slowly over large evolutionary distances, as a result proteins from even evolutionarily distant organisms can be aligned against raw genomic sequence to try and identify regions of homology. MAKER does this using BLASTX.

Remember now that we are aligning against the repeat-masked genomic sequence. How is this going to affect our alignments? For one thing we won't be able to align against low-complexity regions. Some real proteins contain low-complexity regions and it would be nice to identify those, but if I let anything align to a low-complexity region, then I will get spurious alignments all over the genome. Wouldn't it be nice if there was a way to allow BLAST to extend alignments through low-complexity regions, but only if there is is already alignment somewhere else? You can do this with soft-masking. If you remember soft-masking is using lower case letters to mask sequence without losing the sequence information. BLAST allows you to use soft-masking to keep alignments from seeding in low-complexity regions, but allows you to extend through them. This of course will allow some of the spurious alignments you were trying to avoid, but overall you still end up suppressing the majority of poor alignments while letting through enough real alignments to justify the cost.

Polishing Evidence Alignments

Because of oddities associated with how BLAST statistics work, BLAST alignments are not as informative as they could be. BLAST will align regions any where it can, even if the algorithm aligns regions out of order, with multiple overlapping alignments in the exact same region, or with slight overhangs around splice sites.

To get more informative alignments MAKER uses the program Exonerate to polish BLAST hits. Exonerate realigns each sequences identified by BLAST around splice sites and forces the alignments to occur in order. The result is a high quality alignment that can be used to suggest near exact intron/exon positions. Polished alignments are produced using the est2genome and protein2genome options for Exonerate.

One of the benefits of polishing EST alignments is the ability to identify the strand an EST derives from. Because of amplification steps involved in building an EST library and limitations involved in some high throughput sequencing technologies, you don't necessarily know whether you're really aligning the forward or reverse transcript of an mRNA. However, if you take splice sites into account, you can only align to one strand correctly.

Integrating Evidence to Synthesize Final Annotations

Once you have ab initio predictions, EST alignments, and protein alignments you can integrate this evidence to produce even better gene predictions. MAKER does this by "talking" to the gene prediction programs. MAKER takes all the evidence, generates "hints" to where splice sites and protein coding regions are located, and then passes these "hints" to programs that will accept them.

MAKER produces hint based predictors for:

  • SNAP
  • Augustus
  • GeneMark (under development)

MAKER then takes the entire pool of ab initio and evidence informed gene predictions, updates features such as 5' and 3' UTRs based on EST evidence, tries to determine alternative splice forms where EST data permits, produces quality control metrics for each gene model (this is included in the output), and then MAKER chooses from among all the gene model possibilities the one that best matches the evidence. This is done using a modified sensitivity/specificity distance metric.

Running MWAS with your Own Data

When using your own data, you need to tell MWAS all the details about how you want the annotation process to proceed. Because there can be many variables and options involved in annotation you will need to review each option carefully. At the very least you should provide a genome sequence file, an EST sequence file, and a protein homology sequence file for new annotation jobs.

MWAS Job Configuration

Basic Input Files

All the basic input files for MWAS should be in fasta format.

  • genome - Genomic sequence file
  • est - ESTs from the same organism or from a very very closely related organism (i.e. chimpanzee to human). These are aligned first via BLASTN with very strict filtering so any sequence divergence can prohibit the alignment.
  • altest - These are ESTs from other closely related organisms (i.e. mouse to human). They are aligned via TBLASTX in protein space, so greater sequence divergence is permitted.
  • protein - proteins from the same or other organisms. These are aligned via BLASTX against the genome. Proteins that align to a region will not necessarily be orthologous or paralogous. The alignment may just be based on short regions such as a shared domain. You may also get alignments to pseudogenes. Polishing BLASTX hits with Exonerate helps identify what are likely true paralogs and orthologs.

Repeat Masking Options

Repeat masking is important for improving gene predictor performance and avoiding protein alignments to what are likely just transposons. You also expect a certain amount of genomic contamination in the EST database, much of this contamination maps back to repeat regions. By repeat masking we can avoid issues with all types of input data.

  • RepeatMasker - Performs repeat masking using the RepBase libraries.
  • RepeatRunner - This is a fasta file of transposon and virus related proteins. The serve provides an internal database to use by default.
  • Users can also supply a fasta file of species specific nucleotide repeats or a GFF3 file of pre-defined repeat regions. Species specific repeat database can be built using programs like PILER and uploaded for use with MAKER.

Gene Prediction Options

Gene prediction options affect the final gene annotations more than any other option type. This brings up the point that electronically produced gene annotations will only be as good as the gene predictions they are based on.

  • Predictor Options - Tell MWAS which programs to use when generating gene models.
    • SNAP
    • Augustus
    • GeneMark
    • Est2Genome - Allows high quality spliced Exonerate EST alignments to become gene annotations. This only happens when there is no gene prediction overlapping the region. This is useful for generating gene annotations in the absence of a trained gene predictor.
    • Protein2Genome - Used only for Prokaryotic genomes. Will try and build gene models based solely on the presence of open reading frames and protein alignments to other species.
    • User supplied gene predictions - These are gene predictions in GFF3 format from any source you have available to you. They will be treated the same as any gene predictions derived from MWAS supported sources.
    • User supplied gene models - These are pre-existing gene models from the same assembly as the contigs being annotated. They can be integrated and automatically updated by MAKER to reflect new evidence (i.e. add UTR etc.). MAKER can also pull names forward from these pre-existing gene models onto new updated genome annotations.

Other MAKER Options

  • Sets the minimum length a contig must have or else it will be skipped.
  • Sets the minimum length a predicted protein must have (in amino acids) to be annotated.
  • Set the expected max intron size for evidence alignments
  • Tells MAKER to consider single exon EST evidence when generating annotations. Single exon ESTs are more likely to be genomic contamination.
  • 'Sets the minimum length required for single exon ESTs if 'single_exon' is enabled

MWAS Results

The results provided to the user from the MWAS can either be downloaded or directly viewed online using a Java Web Start version of the Apollo genome annotation curration tool.

If you choose to download your data you will be presented with a tarball that when unpacked will produce an output directory called something like 2434.maker.output. The name of the output directory is based off of the job id assigned to your sequence file.

When you examine the contents of this directory, you should see a list of directories and files created by MAKER.

drwxr-xr-x 3 gmod gmod 4096 2009-07-12 23:23 2434_datastore
-rw-r--r-- 1 gmod gmod  135 2009-07-12 23:27 2434_master_datastore_index.log
-rw-r--r-- 1 gmod gmod 1579 2009-07-12 23:23 maker_bopts.log
-rw-r--r-- 1 gmod gmod 1250 2009-07-12 23:23 maker_exe.log
-rw-r--r-- 1 gmod gmod 4016 2009-07-12 23:23 maker_opts.log
drwxr-xr-x 2 gmod gmod 4096 2009-07-12 23:23 mpi_blastdb
  • The maker_opt.log, maker_exe.log, and maker_bopts.log files are logs of the control files used for this run of MAKER.
  • The mpi_blastdb directory contains fasta indexes and BLAST database files created from the input EST, protein, and repeat databases.
  • The 2434_master_datastore_index.log contains information on both the run status of individual contigs and information on where individual contig data is stored.
  • The 2434_datastore directory contains a set of subfolders, each containing the final MAKER output for individual contigs from the genomic fasta file.

Once a MAKER run is finished the most important file to look at is the 2434_master_datastore_index.log to see if there were any failures.

less 2434_master_datastore_index.log.  MWAS provides a summery of this file when you click on results to download a job.  MWAS also displays run errors in the log option button that you can click on when in the MWAS main queue page.

If everything proceeded correctly you should see the following in your 2434_master_datastore_index.log file.

contig-dpp-500-500      2434_datastore/contig-dpp-500-500 STARTED
contig-dpp-500-500      2434_datastore/contig-dpp-500-500 FINISHED

There are only entries describing a single contig because there was only one contig in the example file. These lines indicate that the contig 'contig-dpp-500-500' STARTED and then FINISHED without incident. Other possible entries include:

  • DIED - indicates a failed run on this contig, MAKER will retry these
  • RETRY - indicates that MAKER is retrying a contig that failed
  • SKIPPED_SMALL - indicates the contig was too short
  • DIED_SKIPPED_PERMANENT - indicates a failed contig that MAKER will not attempt to retry

The entries in the 2434_master_datastore_index.log file also indicate that the output files for this contig are stored in the directory dpp_contig_datastore/contig-dpp-500-500/. Knowing where the output is stored may seem rather trivial; however, input genome fasta files can contain thousands even hundreds-of-thousands of contigs, and many file-systems have performance problems with large numbers of sub-directories and files within a single directory. Even when the underlying file-systems handle things gracefully, access via network file-systems can be an issue. To deal with this situation, MAKER uses a datastore module to create a hierarchy of sub-directory layers, starting from a 'base', and mapping identifiers to corresponding sub-directories. For situations where the input genome fasta file contains more than 1,000 contigs, the datastore structure is used automatically, and the master_datastore_index.log file becomes essential for identifying where the output for a given contig is stored.

now let's take a look at what MAKER produced for the contig 'contig-dpp-500-500'.

cd 2434_datastore/contig-dpp-500-500
ls -l

The directory should contain a number of files.

-rw-r--r-- 1 gmod gmod 47437 2009-07-12 23:27 contig-dpp-500-500.gff
-rw-r--r-- 1 gmod gmod   189 2009-07-12 23:27 contig-dpp-500-500.maker.non_overlapping_ab_initio.proteins.fasta
-rw-r--r-- 1 gmod gmod   399 2009-07-12 23:27 contig-dpp-500-500.maker.non_overlapping_ab_initio.transcripts.fasta
-rw-r--r-- 1 gmod gmod   704 2009-07-12 23:27 contig-dpp-500-500.maker.proteins.fasta
-rw-r--r-- 1 gmod gmod   901 2009-07-12 23:27 contig-dpp-500-500.maker.snap_masked.proteins.fasta
-rw-r--r-- 1 gmod gmod  4837 2009-07-12 23:27 contig-dpp-500-500.maker.snap_masked.transcripts.fasta
-rw-r--r-- 1 gmod gmod  4430 2009-07-12 23:27 contig-dpp-500-500.maker.transcripts.fasta

  • The contig-dpp-500-500.gff contains all annotations and evidence alignments in GFF3 format. This is the important file for use with Apollo or GBrowse.
  • The contig-dpp-500-500.maker.transcripts.fasta and contig-dpp-500-500.maker.proteins.fasta files contain the transcript and protein sequences for MAKER produced gene annotations.
  • The contig-dpp-500-500.maker.snap_masked.transcripts.fasta and contig-dpp-500-500.maker.snap_masked.proteins.fasta files contain the transcript and protein sequences for all SNAP ab initio gene predictions. If you use other ab initio gene predictors, those sequence files will follow a similar naming pattern.
  • The contig-dpp-500-500.maker.non_overlapping_ab_initio.transcripts.fasta and contig-dpp-500-500.maker.non_overlapping_ab_initio.proteins.fasta files contain the set of best ab initio gene predictions that do not overlap a MAKER gene annotation. These files can be analyzed to see if there is any reason to promote them to the status of gene annotations. For example: you can run iprscan to see if they contain known protein domains.

Viewing MAKER Annotations

Viewing the raw GFF3 file produced by MAKER really isn't that meaningful.

For sanity checking purposes it would be nice to have a graphical view of what's in the GFF3 file. To do this GFF3 files can be loaded into programs like Apollo and GBrowse. MWAS allows you to view the files in Apollo directly on the website. You can also get summery statistics of annotation features using the tool SOBA from the Sequence Ontology Consortium.


On the results screen choose a contig from a job and click "View in Apollo". A Java Web Start version of Apollo will then install itself automatically on your computer, if not already installed. Apollo will then automatically load the contig you indicated into the browser. You will notice that there are a number of bars representing the gene annotations and the evidence alignments supporting those annotations. Annotations are in the middle light colored panel, and evidence alignments are in the dark panels at the top and bottom.

All the evidence in the dark panels will be a different color depending on the source each piece of evidence was derived from (i.e. RepeatMasker, BLASTX, etc.). To identify which source a feature belongs to, just manually clicking on one and the name of the source will be displayed in the table at the bottom of the Apollo screen.

Possible Sources Include:

  • BLASTN - BLASTN alignment of EST evidence
  • BLASTX - BLASTX alignment of protein evidence
  • TBLASTX - TBLASTX alignment of EST evidence from closely related organisms
  • EST2Genome - Polished EST alignment from Exonerate
  • Protein2Genome - Polished protein alignment from Exonerate
  • SNAP - SNAP ab inito gene prediction
  • GENEMARK - GeneMarkab inito gene prediction
  • Augustus - Augustus ab inito gene prediction
  • FgenesH - FGENESH ab inito gene prediction
  • Repeatmasker - RepeatMasker identified repeat
  • RepeatRunner - RepeatRunner identified repeat from the repeat protein database

Training ab initio Gene Predictors

If you are involved in a genome project for an emerging model organism, you should already have an EST database which would have been generated as part of the original sequencing project. A protein database can be collected from closely related organism genome databases or by using the UniProt/SwissProt protein database or the NCBI NR protein database. However a trained ab initio gene predictor is a much more difficult thing to generate. Gene predictors require existing gene models on which to base prediction parameters. However, with emerging model organisms there are no pre-existing gene models. So how then are you supposed to train your gene prediction programs?

MWAS gives the user the option to produce gene annotations directly from the EST evidence. You can then use these imperfect gene models to train gene predictor program. Once you have re-run MWAS with the newly trained gene predictor, you can use the second set of gene annotations to train the gene predictors yet again. This boot-strap process allows you to iteratively improve the performance of ab initio gene predictors.

GFF3 Pass-through

What if I'm not working on a new genome project, but rather I have an existing annotation set, and I just want to update my genome database to reflect new protein and EST evidence. Here you can use a feature in MAKER called GFF3 pass-through, which allows you to pass existing annotations into the program and combine them with new evidence for use in the annotation process.


mRNAseq is a high throughput technique for sequencing the entire transcriptome, and it holds the promise of allowing researchers to identify all exons and alternative splice forms for every gene in the genome with a single experiment. It may soon make gene predictors (mostly) a thing of the past.

  • Still need to de-convolute reads & evidence (for now)
  • Still need to archive, manage, and distribute annotations


We are currently working on native support for mRNAseq data within the MAKER pipeline. However, because of the GFF3 pass-through option, there is a way to take advantage of mRNAseq reads right now. By mapping mRNAseq reads using BowTie and TopHat, you can create GFF3 files of read islands and junctions. This data can then be passed in as EST evidence and will be used for generating hint based gene prediction and for choosing final annotations.

Merge/Resolve Legacy Annotations

Legacy annotations

  • Many are no longer maintained by original creators
  • In some cases more than one group has annotated the same genome, using very different procedures, even different assemblies
  • Many investigators have their own genome-scale data and would like a private set of annotations that reflect these data
  • There will be a need to revise, merge, evaluate, and verify legacy annotation sets in light of RNA-seq and other data


MAKER will:

  • Identify legacy annotation most consistent with new data
  • Automatically revise it in light of new data
  • If no existing annotation, create new one