GETA

GETA is an automatic genome-wide annotation tool (GWAT) with improved accuracy and gene integrity for eukaryotes written by Lianfu Chen, Congcong Liu, Lei Deng and so on.

1. introduction

GETA can integrate the various evidence from ab initio gene finding and experimental data, including homologous proteins and RNA-Seq data.

GETA is a multi-threaded, automatic, and memory-saving GWAT, which effectively improved the accuracy and completeness of predicted annotation. The current version is v1.0 and later version will support more types of data like the single-molecular sequencing data and add other analysis modules on the basis of continuous improvement in accuracy and completeness.

2. installation

1.unpack

2.install dependencies:

ParaFly
java 1.8.0_144
HISAT2 2.1.0
samtools 1.3.1
hmmer 3.1b2
Augustus

3.add these directories of the executables to the PATH environment variable

3. the usage of the main script geta.pl

Usage:

geta.pl [options]

For example:

geta.pl --RM_species Embryophyta --out_prefix out -1 reads.1 fastq -2 reads.2.fastq --protein homolog.fasta --cpu 80 --hisat2 " --min-intronlen 20 --max-intronlen 20000 --rna-strandness RF" --strand_specific --sam2transfrag " --fraction --min_expressed_base_depth 2 --max_expressed_base_depth 50 --min_junction_depth 2 --max_junction_depth 50" --species oryza_sativa_20171120 --pfam_db /opt/biosoft/hmmer-3.1b2/Pfam-AB.hmm --gene_prefix OS01Gene --genome genome.fasta

Parameters:

--RM_species <string>    default: None
species identifier for RepeatMasker.

--genome <string>
genome file in fasta format.

--out_prefix <string>    default: out
the prefix of outputs.

-1 <string> -2 <string>
fastq format files contain of paired-end RNA-seq data.

-S <string>
fastq format file contains of single-end RNA-seq data.

--protein <string>
homologous protein sequences (derived from multiple species would be recommended) file in fasta format.

--cpu <int>    default: 4
the number of threads.

--trimmomatic <string>    default: "TruSeq3-PE-2.fa:2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:50 TOPHRED33"
the parameter of trimmomatic. The value must be surrounded by a pair of quotes.

--hisat2 <int>    default: " --min-intronlen 20 --max-intronlen 20000 --rna-strandness unstranded --dta --score-min L,0.0,-0.4"
the parameter of hisat2. The value must be surrounded by a pair of quotes, and the first quote is followed by a white space.

--strand_specific
enable the ability of analysing the strand-specific information provided by the tag "XS" of alignments. If this parameter was set, the paramter "--rna-strandness" should be set to "RF" usually, so the value of "--hisat2" would be " --min-intronlen 20 --max-intronlen 20000 --rna-strandness RF --dta --score-min L,0.0,-04".

--sam2transfrag <string>    default: " --fraction 0.05 --min_expressed_base_depth 2 --max_expressed_base_depth 50 --min_junction_depth 2 --max_junction_depth 50"
the parameter of sam2transfrag. The value must be surrounded by a pair of quotes, and the first quote is followed by a white space.
the defalut value means: the dynamic coverage threshold of mapping regions (come frome SAM file) was determinated by the maximum base depth of each region * 0.05, as well as this threshold should between 2 and 50.

--ORF2bestGeneModels <string>    default: " --min_cds_num 3 --min_cds_length 900 --min_cds_exon_ratio 0.60 --intron_length_fractile 0.95 --cds_length_fractile 0.95"
the parameter of ORF2bestGeneModels. The value must be surrounded by a pair of quotes, and the first quote is followed by a white space.
the defalut value means: a good gene model shold be completed, contain at least 3 CDS sequences whose total length shold >= 900bp, the length ratio of CDS/exon should >= 60%, intron and cds length should not great than a auto-calculated threashold.

--species <string>
species identifier for Augustus.  

--pfam_db <string>    default: "/opt/biosoft/hmmer-3.1b2/Pfam-AB.hmm"
the absolute path of Pfam database which was used for filtering of false positive gene models.

--gene_prefix <string>    default: gene
the prefix of gene id shown in output file.

This script was tested on CentOS 6.8 with such softwares can be run directly in terminal:

1. ParaFly
2. java version "1.8.0_144"
3. HISAT2 version 2.1.0
4. samtools Version: 1.3.1
5. hmmer-3.1b2

Version: 1.0

4. testing of GETA

we have compared the accuracy of GETA with other methods according to several species. the command line and output GFF3 files can be found at website: 122.205.95.116/geta/