ORForise - Platform for analysing and comparing Prokaryote CoDing Sequence (CDS) Gene Predictions.
Project description
ORForise - Prokaryote Genome Annotation Analysis and Comparison Platform
Published in Bioinformatics : https://academic.oup.com/bioinformatics/article/38/5/1198/6454948
Platform for analysing and comparing Prokaryote CoDing Sequence (CDS) Gene Predictions.
Novel genome annotations can be compared to a provided reference annotation from Ensembl and predictions from other tools (or any given GFF annotation) .
Requirements and Installation:
The ORForise platform is written in Python (3.6-3.9) and only requires the NumPy library (should be installed automatically by pip when installing ORForise) which is standard in most base installations of Python3.
Intallation:
The ORForise platform is available via the pip Python package manager pip3 install ORForise
.
Consider using '--no-cache-dir' with pip to ensure the download of the newest version of the package.
Required Files:
To run, you need:
- Input Genome FASTA and corresponding GFF file (or CDS predictions with the annotated genes for the genome you want to use as reference in one of the tool output formats listed below).
- A prediction output from one of the compatible tools for the same genome.
How to add your own Genome:
Corresponding FASTA and GFF files must be provided for the genome the analysis is to be performed on, including the corresponding output of any tools to compare.
How to add your own tool:
If the new tool reports its predictions in GFF you can present ORForise with "GFF" for either the reference -rt
or prediction -t
option.
If the tool uses another non-standard format, a request can be made to add it as an option via GitHub.
Testing:
Precomputed testing and data which includes example input and output files for all tools presented below is available in the ~ORForise/Testing
directory of the GitHub repository.
Example output files from Annotation-Compare
, GFF-Adder
and GFF-Intersector
are made available to validate installation.
CDS Prediction Analysis:
Use-cases: (Running if via pip)
For Help: Annotation-Compare -h
Thank you for using ORForise
Please report any issues to: https://github.com/NickJD/ORForise/issues
#####
usage: Annotation_Compare.py [-h] -dna GENOME_DNA -ref REFERENCE_ANNOTATION -t TOOL -tp TOOL_PREDICTION
[-rt REFERENCE_TOOL] [-o OUTNAME] [-v {True,False}]
ORForise v1.4.2: Annotatione-Compare Run Parameters.
Required Arguments:
-dna GENOME_DNA Genome DNA file (.fa) which both annotations are based on
-ref REFERENCE_ANNOTATION
Which reference annotation file to use as reference?
-t TOOL Which tool to analyse? (Prodigal)
-tp TOOL_PREDICTION Tool genome prediction file (.gff) - Different Tool Parameters are compared individually via
separate files
Optional Arguments:
-rt REFERENCE_TOOL What type of Annotation to compare to? -- Leave blank for Ensembl reference- Provide tool
name to compare output from two tools
Output:
-o OUTNAME Define full output filename (format is CSV) - If not provided, summary will be printed to
std-out
Misc:
-v {True,False} Default - False: Print out runtime status
Compare a novel genome annotation to an Ensembl annotation:
Genome annotation is a difficult process, even for Prokaryotes. ORForise allows the direct and systematic analysis of a novel CDS prediction from a wide selection of tools to a reference Genome Annotation, such as those provided by Ensembl Bacteria.
Example: Installation through pip will allow user to call the programs directly from the ORForise package.
Annotation-Compare -dna ~/Testing/Myco.fa -ref ~/Testing/Myco.gff -t Prodigal -tp ~/Testing/Prodigal_Myco.gff
Compare different novel annotations with each other on a single Genome:
If a reference Genome Annotation is not available or a direct comparison between two or more tools is wanted, ORForise can be used as the example below.
Aggregate CDS Prediction Analysis:
Use-cases: (Running if via pip)
For Help: Aggregate-Compare -h
Thank you for using ORForise
Please report any issues to: https://github.com/NickJD/ORForise/issues
#####
usage: Aggregate_Compare.py [-h] -dna GENOME_DNA -t TOOLS -tp TOOL_PREDICTIONS -ref REFERENCE_ANNOTATION
[-rt REFERENCE_TOOL] [-o OUTNAME] [-v {True,False}]
ORForise v1.4.2: Aggregate-Compare Run Parameters.
Required Arguments:
-dna GENOME_DNA Genome DNA file (.fa) which both annotations are based on
-t TOOLS Which tools to analyse? (Prodigal,GeneMarkS)
-tp TOOL_PREDICTIONS Tool genome prediction file (.gff) - Providefile locations for each tool comma separated
-ref REFERENCE_ANNOTATION
Which reference annotation file to use as reference?
Optional Arguments:
-rt REFERENCE_TOOL What type of Annotation to compare to? -- Leave blank for Ensembl reference- Provide tool
name to compare output from two tools
Output:
-o OUTNAME Define full output filename (format is CSV) - If not provided, summary will be printed to
std-out
Misc:
-v {True,False} Default - False: Print out runtime status
Example:
Aggregate-Compare -ref ~/Testing/Myco.gff -dna ~/Testing/Myco.fa -t Prodigal,TransDecoder,GeneMark_S_2 -tp ~/Testing/Prodigal_Myco.gff,~/Testing/TransDecoder_Myco.gff,~/Testing/GeneMark_S_2_Myco.gff
This will compare the Aggregate the predictions of Prodigal, TransDecoder and GLIMMER 3 against the Mycoplasma reference annotation provided by Ensembl Bacteria.
Annotation Comparison Output - The output format is the same for Annotation_Compare and Aggregate_Compare:
Print to screen example - Prodigal prediction compared to Ensembl Bacteria reference annotation of Escherichia coli:
Annotation-Compare.py -ref ./Testing/Myco.gff -dna ./Testing/Myco.fa -t Prodigal -tp ./Testing/Prodigal_Myco.gff
Genome Used: Myco
Reference Used: Testing/Myco.gff
Tool Compared: Prodigal
Perfect Matches:128[476] -26.89%
Partial Matches:62[476] - 13.03%
Missed Genes:286[476] - 60.08%
Complete
Aggregate-Compare -ref ./Testing/Myco.gff -dna ./Testing/Myco.fa -t Prodigal,TransDecoder,GeneMark_S_2 -tp ./Testing/Prodigal_Myco.gff,./Testing/TransDecoder_Myco.gff,./Testing/GeneMark_S_2_Myco.gff
Prodigal
TransDecoder
GeneMark_S_2
Match filtered out
Match filtered out
Match filtered out
Match filtered out
Match filtered out
Match filtered out
Genome Used: Myco
Reference Used: ./Testing/Myco.gff
Tools Compared: Prodigal,TransDecoder,GeneMark_S_2
Perfect Matches:132[476]
Partial Matches:58[476]
Missed Genes:286[476]
This is the default output of the comparison tools.
'-o' Example output to CSV file - Prodigal prediction compared to Ensembl Bacteria reference annotation of Escherichia coli:
The output is designed to be human-readable and interpretable by the included 'ORForise_Analysis' scripts. The example below presents the 12 'Representative' and 72 'All' Metrics but only shows one entry for each of the induvidual prediction reports (Perfect_Match_Genes,Partial_Match_Genes,Missed_Genes,Predicted_CDS_Without_Corresponding_Gene_in_Reference,Predicted_CDSs_Which_Detected_more_than_one_Gene).
Representative_Metrics:
Percentage_of_Genes_Detected,Percentage_of_ORFs_that_Detected_a_Gene,Percent_Difference_of_All_ORFs,Median_Length_Difference,Percentage_of_Perfect_Matches,Median_Start_Difference_of_Matched_ORFs,Median_Stop_Difference_of_Matched_ORFs,Percentage_Difference_of_Matched_Overlapping_CDSs,Percent_Difference_of_Short-Matched-ORFs,Precision,Recall,False_Discovery_Rate
39.92,19.10,109.03,-62.17,67.37,67.5,-85.5,-83.71,-17.39,0.19,0.40,0.81
All_Metrics:
Number_of_ORFs,Percent_Difference_of_All_ORFs,Number_of_ORFs_that_Detected_a_Gene,Percentage_of_ORFs_that_Detected_a_Gene,Number_of_Genes_Detected,Percentage_of_Genes_Detected,Median_Length_of_All_ORFs,Median_Length_Difference,Minimum_Length_of_All_ORFs,Minimum_Length_Difference,Maximum_Length_of_All_ORFs,Maximum_Length_Difference,Median_GC_content_of_All_ORFs,Percent_Difference_of_All_ORFs_Median_GC,Median_GC_content_of_Matched_ORFs,Percent_Difference_of_Matched_ORF_GC,Number_of_ORFs_which_Overlap_Another_ORF,Percent_Difference_of_Overlapping_ORFs,Maximum_ORF_Overlap,Median_ORF_Overlap,Number_of_Matched_ORFs_Overlapping_Another_ORF,Percentage_Difference_of_Matched_Overlapping_CDSs,Maximum_Matched_ORF_Overlap,Median_Matched_ORF_Overlap,Number_of_Short-ORFs,Percent_Difference_of_Short-ORFs,Number_of_Short-Matched-ORFs,Percent_Difference_of_Short-Matched-ORFs,Number_of_Perfect_Matches,Percentage_of_Perfect_Matches,Number_of_Perfect_Starts,Percentage_of_Perfect_Starts,Number_of_Perfect_Stops,Percentage_of_Perfect_Stops,Number_of_Out_of_Frame_ORFs,Number_of_Matched_ORFs_Extending_a_Coding_Region,Percentage_of_Matched_ORFs_Extending_a_Coding_Region,Number_of_Matched_ORFs_Extending_Start_Region,Percentage_of_Matched_ORFs_Extending_Start_Region,Number_of_Matched_ORFs_Extending_Stop_Region,Percentage_of_Matched_ORFs_Extending_Stop_Region,Number_of_All_ORFs_on_Positive_Strand,Percentage_of_All_ORFs_on_Positive_Strand,Number_of_All_ORFs_on_Negative_Strand,Percentage_of_All_ORFs_on_Negative_Strand,Median_Start_Difference_of_Matched_ORFs,Median_Stop_Difference_of_Matched_ORFs,ATG_Start_Percentage,GTG_Start_Percentage,TTG_Start_Percentage,ATT_Start_Percentage,CTG_Start_Percentage,Other_Start_Codon_Percentage,TAG_Stop_Percentage,TAA_Stop_Percentage,TGA_Stop_Percentage,Other_Stop_Codon_Percentage,True_Positive,False_Positive,False_Negative,Precision,Recall,False_Discovery_Rate,Nucleotide_True_Positive,Nucleotide_False_Positive,Nucleotide_True_Negative,Nucleotide_False_Negative,Nucleotide_Precision,Nucleotide_Recall,Nucleotide_False_Discovery_Rate,ORF_Nucleotide_Coverage_of_Genome,Matched_ORF_Nucleotide_Coverage_of_Genome
995,109.03,190,19.10,190,39.92,335.0,-62.17,89,-21.24,3152,-41.81,31.50,0.20,32.83,4.42,279,26.24,135,0.00,36,-83.71,31,4.50,443,1826.09,19,-17.39,128,67.37,162,85.26,154,81.05,0,0,0.00,4,2.11,0,0.00,570,0.57,425,0.43,67.5,-85.5,63.12,15.28,21.61,0.00,0.00,0.00,11.06,27.44,61.51,0.00,0.40,1.69,0.60,0.19,0.40,0.81,0.82,0.31,0.69,0.18,0.96,0.82,0.04,77.15,24.47
CDS_Gene_Coverage_of_Genome:
90.62
Start_Position_Difference:
-78,33,93,294,144,408,3,18,156,-42,45,90,333,333,-39,111,201,93,120,-354,-150,-366,117,-138,-240,123,-153,-51
Stop_Position_Difference:
-192,-147,108,-216,87,-678,-96,-156,-321,-240,-168,-162,-51,-126,-33,-3,-93,-12,-204,-189,-156,237,-45,-219,-201,-537,-30,-78,159,243,60,21,15,183,288,6
Alternative_Starts_Predicted:
Alternative_Stops_Predicted:
Undetected_Gene_Metrics:
ATG_Start ,GTG_Start ,TTG_Start ,ATT_Start ,CTG_Start ,Alternative_Start_Codon ,TGA_Stop ,TAA_Stop ,TAG_Stop ,Alternative_Stop_Codon ,Median_Length ,ORFs_on_Positive_Strand ,ORFs_on_Negative_Strand
88.46,7.69,3.85,0.00,0.00,0.00,0.00,74.13,25.87,0.00,1047.50,156,130
Perfect_Match_Genes:
>Myco_686_1828_+
ATGAAAATATTAATTAATAAAAGTGAATTGAATAAAATTTTGAAAAAAATGAATAACGTTATTATTTCCAATAACAAAATAAAACCACATCATTCATATTTTTTAATAGAGGCAAAAGAAAAAGAAATAAACTTTTATGCTAACAATGAATACTTTTCTGTCAAATGTAATTTAAATAAAAATATTGATATTCTTGAACAAGGCTCCTTAATTGTTAAAGGAAAAATTTTTAACGATCTTATTAATGGCATAAAAGAAGAGATTATTACTATTCAAGAAAAAGATCAAACACTTTTGGTTAAAACAAAAAAAACAAGTATTAATTTAAACACAATTAATGTGAATGAATTTCCAAGAATAAGGTTTAATGAAAAAAACGATTTAAGTGAATTTAATCAATTCAAAATAAATTATTCACTTTTAGTAAAAGGCATTAAAAAAATTTTTCACTCAGTTTCAAATAATCGTGAAATATCTTCTAAATTTAATGGAGTAAATTTCAATGGATCCAATGGAAAAGAAATATTTTTAGAAGCTTCTGACACTTATAAACTATCTGTTTTTGAGATAAAGCAAGAAACAGAACCATTTGATTTCATTTTGGAGAGTAATTTACTTAGTTTCATTAATTCTTTTAATCCTGAAGAAGATAAATCTATTGTTTTTTATTACAGAAAAGATAATAAAGATAGCTTTAGTACAGAAATGTTGATTTCAATGGATAACTTTATGATTAGTTACACATCGGTTAATGAAAAATTTCCAGAGGTAAACTACTTTTTTGAATTTGAACCTGAAACTAAAATAGTTGTTCAAAAAAATGAATTAAAAGATGCACTTCAAAGAATTCAAACTTTGGCTCAAAATGAAAGAACTTTTTTATGCGATATGCAAATTAACAGTTCTGAATTAAAAATAAGAGCTATTGTTAATAATATCGGAAATTCTCTTGAGGAAATTTCTTGTCTTAAATTTGAAGGTTATAAACTTAATATTTCTTTTAACCCAAGTTCTCTATTAGATCACATAGAGTCTTTTGAATCAAATGAAATAAATTTTGATTTCCAAGGAAATAGTAAGTATTTTTTGATAACCTCTAAAAGTGAACCTGAACTTAAGCAAATATTGGTTCCTTCAAGATAA
>Myco_4812_7322_+
ATGGCAAAGCAACAAGATCAAGTAGATAAGATTCGTGAAAACTTAGACAATTCAACTGTCAAAAGTATTTCATTAGCAAATGAACTTGAGCGTTCATTCATGGAATATGCTATGTCAGTTATTGTTGCTCGTGCTTTACCTGATGCTAGAGATGGACTTAAACCAGTTCATCGTCGTGTTCTTTATGGTGCTTATATTGGTGGCATGCACCATGATCGTCCTTTTAAAAAGTCTGCGAGGATTGTTGGTGATGTAATGAGTAAATTCCACCCTCATGGTGATATGGCAATATATGACACCATGTCAAGAATGGCTCAAGACTTTTCATTAAGATACCTTTTAATTGATGGTCATGGTAATTTTGGTTCTATAGATGGTGATAGACCTGCTGCACAACGTTATACAGAAGCAAGATTATCTAAACTTGCAGCAGAACTTTTAAAAGATATTGATAAAGATACAGTTGACTTTATTGCTAATTATGATGGTGAGGAAAAAGAACCAACTGTTCTACCAGCAGCTTTCCCTAACTTACTTGCAAATGGTTCTAGTGGGATTGCAGTTGGAATGTCAACATCTATTCCTTCCCATAATCTCTCTGAATTAATTGCGGGTTTAATCATGTTAATTGATAATCCTCAATGCACTTTTCAAGAATTATTAACTGTAATTAAAGGACCTGATTTTCCAACAGGAGCTAACATTATCTACACAAAAGGAATTGAAAGCTACTTTGAAACAGGTAAAGGCAATGTAGTAATTCGTTCTAAAGTTGAGATAGAACAATTGCAAACAAGAAGTGCATTAGTTGTAACTGAAATTCCTTACATGGTTAACAAAACTACCTTAATTGAAAAGATTGTAGAACTTGTTAAAGCTGAAGAGATTTCAGGAATTGCTGATATCCGTGATGAATCCTCTCGAGAAGGAATAAGGTTAGTGATTGAAGTAAAACGCGACACTGTACCTGAAGTTTTATTAAATCAACTTTTTAAATCAACAAGATTACAAGTACGCTTCCCTGTTAATATGCTTGCTTTAGTTAAAGGAGCTCCTGTACTTCTCAACATGAAACAAGCTTTGGAAGTATATCTTGATCATCAAATTGATGTTCTTGTTAGAAAAACAAAGTTTGTGCTTAATAAACAACAAGAACGTTATCACATTTTAAGCGGACTTTTAATTGCTGCTTTAAATATTGATGAGGTTGTTGCAATTATTAAAAAATCAGCAAATAACCAGGAAGCAATTAATACATTAAATACAAAGTTTAAGCTTGATGAAATTCAAGCTAAAGCAGTTCTTGACATGCGTTTAAGGAGCTTAAGCGTACTTGAAGTTAACAAACTTCAAACTGAACAAAAAGAGTTAAAAGATTCAATTGAATTTTGTAAGAAAGTGTTAGCTGATCAAAAATTACAGCTAAAAATAATCAAAGAGGAATTGCAAAAAATCAATGATCAGTTTGGTGATGAAAGAAGAAGTGAAATTCTCTATGATATCTCTGAGGAAATTGATGATGAATCATTGATAAAAGTTGAGAATGTAGTGATAACTATGTCTACAAATGGTTATCTAAAAAGGATTGGAGTTGATGCTTATAATCTTCAACATCGTGGTGGAGTTGGGGTTAAAGGGCTAACTACTTATGTTGATGATAGTATTAGTCAATTATTGGTCTGTTCAACTCACTCTGACTTATTATTTTTTACTGATAAGGGTAAGGTTTATAGAATTAGAGCTCATCAAATTCCCTATGGTTTTAGAACAAATAAAGGTATTCCCGCTGTTAACTTAATCAAAATTGAAAAGGATGAAAGAATTTGTTCATTGTTATCTGTTAATAACTATGATGATGGTTATTTCTTTTTCTGTACTAAAAATGGAATTGTTAAAAGAACGAGCTTGAATGAATTCATCAACATCTTAAGTAATGGTAAGCGGGCTATATCTTTTGATGATAATGACACTTTGTATTCAGTAATTAAAACCCACGGAAATGATGAGATTTTTATTGGTTCTACCAATGGATTTGTTGTTCGCTTCCATGAAAATCAACTCAGAGTTCTTTCAAGAACAGCAAGAGGTGTATTTGGTATCAGTTTAAATAAAGGAGAATTTGTTAATGGACTATCAACTTCAAGCAACGGTAGCTTACTTTTATCAGTCGGTCAAAATGGAATAGGTAAATTAACGAGCATAGATAAATATAGACTCACAAAACGTAATGCTAAGGGAGTTAAAACTCTAAGGGTTACTGATAGAACAGGCCCTGTTGTTACAACAACCACTGTTTTTGGTAATGAGGATCTTTTAATGATTTCCTCTGCTGGTAAAATTGTGCGTACCAGTTTACAAGAACTTTCAGAACAAGGTAAAAACACTTCTGGTGTTAAGTTAATTAGATTAAAAGATAATGAACGTTTAGAAAGAGTAACTATCTTTAAAGAAGAGTTAGAAGACAAAGAAATGCAACTAGAAGATGTTGGATCCAAACAAATTACGCAATAA
.........
Partial_Match_Genes:
Gene:9923_11251_+_ATG_TAA
ATGAAAAGCGAAATTAATATTTTTGCACTAGCAACTGCACCTTTTAATAGTGCATTACATATTATTAGGTTTTCTGGTCCTGATGTTTATGAGATTTTAAACAAGATAACTAATAAAAAAATAACAAGAAAAGGGATGCAAATTCAACGCACATGGATAGTTGATGAAAACAATAAGCGAATTGATGATGTGCTATTATTTAAATTTGTCTCTCCAAATTCTTATACAGGAGAAGATTTAATTGAAATTTCTTGTCATGGTAACATGTTGATCGTTAATGAAATTTGCGCACTTCTTTTAAAAAAAGGAGGTGTTTATGCCAAACCTGGTGAATTTACCCAAAGGAGTTTTTTAAATGGAAAAATGAGTTTACAACAAGCTAGTGCTGTAAATAAATTGATTTTATCTCCTAACTTATTAGTTAAAGATATAGTCTTAAATAATTTAGCGGGTGAAATGGATCAACAATTAGAACAAATAGCTCAACAAGTTAATCAATTAGTAATGCAAATGGAAGTAAACATTGATTATCCAGAATATCTTGATGAACAAGTAGAACTATCAACTTTAAATAATAAAGTTAAATTGATTATTGAAAAGCTTAAAAGAATTATTGAAAATAGTAAACAACTCAAAAAACTTCACGATCCTTTTAAAATTGCCATTATAGGCGAAACTAATGTAGGTAAATCTTCTTTACTCAACGCTTTATTAAATCAAGATAAAGCGATAGTTTCAAATATTAAAGGTAGTACACGCGATGTTGTTGAAGGGGATTTCAATTTAAATGGTTATTTAATCAAGATCTTAGATACTGCAGGTATCCGTAAACATAAAAGTGGGCTTGAAAAAGCAGGAATTAAAAAAAGCTTTGAATCTATAAAGCAAGCTAATTTGGTTATTTATCTTTTAGATGCAACACATCCAAAGAAAGATCTTGAATTAATTAGTTTTTTTAAGAAAAATAAAAAGGATTTTTTTGTTTTCTATAACAAAAAAGATTTAATTACAAATAAGTTTGAAAATAGTATTTCTGCAAAGCAAAAAGATATTAAAGAATTAGTTGATTTATTAACTAAATATATTAACGAGTTTTATAAAAAAATAGATCAAAAAATCTATCTGATTGAAAATTGACAGCAAATTTTAATTGAAAAAATTAAAGAACAATTAGAACAGTTTTTAAAGCAACAAAAAAAATATTTATTTTTCGATGTTTTAGTTACCCATCTAAGAGAAGCTCAACAAGATATTCTTAAACTACTAGGTAAGGATGTAGGTTTTGATTTAGTTAATGAAATTTTTAATAATTTTTGTTTAGGAAAATAA
ORF:9923_11059_+_ATG_TGA
ATGAAAAGCGAAATTAATATTTTTGCACTAGCAACTGCACCTTTTAATAGTGCATTACATATTATTAGGTTTTCTGGTCCTGATGTTTATGAGATTTTAAACAAGATAACTAATAAAAAAATAACAAGAAAAGGGATGCAAATTCAACGCACATGGATAGTTGATGAAAACAATAAGCGAATTGATGATGTGCTATTATTTAAATTTGTCTCTCCAAATTCTTATACAGGAGAAGATTTAATTGAAATTTCTTGTCATGGTAACATGTTGATCGTTAATGAAATTTGCGCACTTCTTTTAAAAAAAGGAGGTGTTTATGCCAAACCTGGTGAATTTACCCAAAGGAGTTTTTTAAATGGAAAAATGAGTTTACAACAAGCTAGTGCTGTAAATAAATTGATTTTATCTCCTAACTTATTAGTTAAAGATATAGTCTTAAATAATTTAGCGGGTGAAATGGATCAACAATTAGAACAAATAGCTCAACAAGTTAATCAATTAGTAATGCAAATGGAAGTAAACATTGATTATCCAGAATATCTTGATGAACAAGTAGAACTATCAACTTTAAATAATAAAGTTAAATTGATTATTGAAAAGCTTAAAAGAATTATTGAAAATAGTAAACAACTCAAAAAACTTCACGATCCTTTTAAAATTGCCATTATAGGCGAAACTAATGTAGGTAAATCTTCTTTACTCAACGCTTTATTAAATCAAGATAAAGCGATAGTTTCAAATATTAAAGGTAGTACACGCGATGTTGTTGAAGGGGATTTCAATTTAAATGGTTATTTAATCAAGATCTTAGATACTGCAGGTATCCGTAAACATAAAAGTGGGCTTGAAAAAGCAGGAATTAAAAAAAGCTTTGAATCTATAAAGCAAGCTAATTTGGTTATTTATCTTTTAGATGCAACACATCCAAAGAAAGATCTTGAATTAATTAGTTTTTTTAAGAAAAATAAAAAGGATTTTTTTGTTTTCTATAACAAAAAAGATTTAATTACAAATAAGTTTGAAAATAGTATTTCTGCAAAGCAAAAAGATATTAAAGAATTAGTTGATTTATTAACTAAATATATTAACGAGTTTTATAAAAAAATAGATCAAAAAATCTATCTGATTGAAAATTGA
Gene:11251_12039_+_ATG_TAA
ATGGAATACTTTGATGCACATTGTCATTTAAATTGTGAACCTTTACTGAGTGAAATTGAAAAAAGCATCGCTAATTTCAAATTAATTAATTTAAAAGCAAATGTTGTAGGTACAGATTTGGATAATTCTAAAATTGCTGTTGAATTAGCTAAAAAATATCCTGATCTTTTAAAAGCAACCATAGGTATCCATCCAAATGATGTTCATTTAGTTGATTTTAAAAAGACAAAAAAACAACTTAATGAACTATTAATAAATAACAGAAATTTCATAAGTTGTATTGGTGAATATGGTTTTGATTATCACTACACAACAGAATTTATTGAATTGCAAAACAAATTCTTTGAGATGCAATTTGAAATAGCTGAAACTAATAAATTGGTTCACATGCTTCATATTCGTGATGCTCATGAAAAAATTTATGAAATATTAACAAGATTAAAGCCAACTCAACCTGTGATTTTTCATTGTTTCAGTCAAGATATAAATATTGCTAAAAAGCTACTATCATTAAAAGATTTAAATATTGACATCTTCTTTTCTATCCCAGGGATAGTTACTTTTAAGAATGCTCAAGCATTACATGAAGCTTTAAAGATTATTCCTAGTGAATTACTTTTAAGTGAAACTGACTCACCGTGATTAACCCCTTCTCCTTTTCGAGGCAAAGTTAACTGACCTGAATATGTAGTTCATACTGTTAGCACTGTTGCTGAAATAAAAAAAATAGAAATTGCTGAAATGAAGCGAATTATTGTTAAAAATGCAAAAAAATTATTTTGACATTAA
ORF:11251_11892_+_ATG_TGA
ATGGAATACTTTGATGCACATTGTCATTTAAATTGTGAACCTTTACTGAGTGAAATTGAAAAAAGCATCGCTAATTTCAAATTAATTAATTTAAAAGCAAATGTTGTAGGTACAGATTTGGATAATTCTAAAATTGCTGTTGAATTAGCTAAAAAATATCCTGATCTTTTAAAAGCAACCATAGGTATCCATCCAAATGATGTTCATTTAGTTGATTTTAAAAAGACAAAAAAACAACTTAATGAACTATTAATAAATAACAGAAATTTCATAAGTTGTATTGGTGAATATGGTTTTGATTATCACTACACAACAGAATTTATTGAATTGCAAAACAAATTCTTTGAGATGCAATTTGAAATAGCTGAAACTAATAAATTGGTTCACATGCTTCATATTCGTGATGCTCATGAAAAAATTTATGAAATATTAACAAGATTAAAGCCAACTCAACCTGTGATTTTTCATTGTTTCAGTCAAGATATAAATATTGCTAAAAAGCTACTATCATTAAAAGATTTAAATATTGACATCTTCTTTTCTATCCCAGGGATAGTTACTTTTAAGAATGCTCAAGCATTACATGAAGCTTTAAAGATTATTCCTAGTGAATTACTTTTAAGTGAAACTGACTCACCGTGA
.......
Missed_Genes:
>Myco_1828_2760_+
ATGAATCTTTACGATCTTTTAGAACTACCAACTACAGCATCAATAAAAGAAATAAAAATTGCTTATAAAAGATTAGCAAAGCGTTATCACCCTGATGTAAATAAATTAGGTTCGCAAACTTTTGTTGAAATTAATAATGCTTATTCAATATTAAGTGATCCTAACCAAAAGGAAAAATATGATTCAATGCTGAAAGTTAATGATTTTCAAAATCGCATCAAAAATTTAGATATTAGTGTTAGATGACATGAAAATTTCATGGAAGAACTCGAACTTCGTAAGAACTGAGAATTTGATTTTTTTTCATCTGATGAAGATTTCTTTTATTCTCCATTTACAAAAAACAAATATGCTTCCTTTTTAGATAAAGATGTTTCTTTAGCTTTTTTTCAGCTTTACAGCAAGGGCAAAATAGATCATCAATTGGAAAAATCTTTATTGAAAAGAAGAGATGTAAAAGAAGCTTGTCAACAGAATAAAAATTTTATTGAAGTTATAAAAGAGCAATATAACTATTTTGGTTGAATTGAAGCTAAGCGTTATTTCAATATTAATGTTGAACTTGAGCTCACACAGAGAGAGATAAGAGATAGAGATGTTGTTAACCTACCTTTAAAAATTAAAGTTATTAATAATGATTTTCCAAATCAACTCTGATATGAAATTTATAAAAACTATTCATTTCGCTTATCTTGAGATATAAAAAATGGTGAAATTGCTGAATTTTTCAATAAAGGTAATAGAGCTTTAGGATGAAAAGGTGACTTAATTGTCAGAATGAAAGTAGTTAATAAAGTAAACAAAAGACTGCGTATTTTTTCAAGCTTTTTTGAGAACGATAAATCTAAATTATGGTTCCTTGTTCCAAACGATAAACAAAGTAATCCTAATAAGGGCGTTTTTAACTATAAAACTCAGCACTTTATTGATTAA
>Myco_2845_4797_+
ATGGAAGAAAATAACAAAGCAAATATCTATGACTCTAGTAGCATTAAGGTCCTTGAAGGACTTGAGGCTGTTAGAAAACGCCCTGGAATGTACATTGGTTCTACTGGCGAAGAAGGTTTGCATCACATGATCTGAGAGATAGTAGACAACTCAATTGATGAAGCAATGGGAGGTTTTGCCAGTTTTGTTAAGCTTACCCTTGAAGATAATTTTGTTACCCGTGTAGAGGATGATGGAAGAGGGATACCTGTTGATATCCATCCTAAGACTAATCGTTCTACAGTTGAAACAGTTTTTACAGTTCTACACGCTGGCGGTAAATTTGATAACGATAGCTATAAAGTGTCAGGTGGTTTACACGGTGTTGGTGCATCAGTTGTTAATGCGCTTAGTTCTTCTTTTAAAGTTTGAGTTTTTCGTCAAAATAAAAAGTATTTTCTCAGCTTTAGCGATGGAGGAAAGGTAATTGGAGATTTGGTCCAAGAAGGTAACTCTGAAAAAGAGCATGGAACAATTGTTGAGTTTGTTCCTGATTTCTCTGTAATGGAAAAGAGTGATTACAAACAAACTGTAATTGTAAGCAGACTCCAGCAATTAGCTTTTTTAAACAAGGGAATAAGAATTGACTTTGTTGATAATCGTAAACAAAACCCACAGTCTTTTTCTTGAAAATATGATGGGGGATTGGTTGAATATATCCACCACCTAAACAACGAAAAAGAACCACTTTTTAATGAAGTTATTGCTGATGAAAAAACTGAAACTGTAAAAGCTGTTAATCGTGATGAAAACTACACAGTAAAGGTTGAAGTTGCTTTTCAATATAACAAAACATACAACCAATCAATTTTCAGTTTTTGTAACAACATTAATACTACAGAAGGTGGAACCCATGTGGAAGGTTTTCGTAATGCACTTGTTAAGATCATTAATCGCTTTGCTGTTGAAAATAAATTCCTAAAAGATAGTGATGAAAAGATTAACCGTGATGATGTTTGTGAAGGATTAACTGCTATTATTTCCATTAAACACCCAAACCCACAATATGAAGGACAAACTAAAAAGAAGTTAGGTAATACTGAGGTAAGACCTTTAGTTAATAGTGTTGTTAGTGAAATCTTTGAACGCTTCATGTTAGAAAACCCACAAGAAGCAAACGCTATCATCAGAAAAACACTTTTAGCTCAAGAAGCGAGAAGAAGAAGTCAAGAGGCTAGGGAGTTAACTCGTCGTAAATCACCTTTTGATAGTGGTTCATTACCAGGTAAATTAGCTGATTGTACAACCAGAGATCCTTCGATTAGTGAACTTTACATTGTTGAGGGTGATAGTGCTGGTGGCACTGCTAAAACAGGAAGAGATCGTTATTTTCAAGCTATCTTACCCTTAAGAGGAAAGATTTTAAACGTTGAAAAATCTAACTTTGAACAAATCTTTAATAATGCAGAAATTTCTGCATTAGTGATGGCAATAGGCTGTGGGATTAAACCTGATTTTGAACTTGAAAAACTTAGATATAGCAAGATTGTGATCATGACAGATGCTGATGTTGATGGTGCACACATAAGAACACTTCTCTTAACTTTCTTTTTTCGCTTTATGTATCCTTTGGTTGAACAAGGCAATATTTTTATTGCTCAACCCCCACTTTATAAAGTGTCATATTCCCATAAGGATTTATACATGCACACTGATGTTCAACTTGAACAGTGAAAAAGTCAAAACCCTAACGTAAAGTTTGGGTTACAAAGATATAAAGGACTTGGAGAAATGGATGCATTGCAGCTGTGAGAAACAACAATGGATCCTAAGGTTAGAACATTGTTAAAAGTTACTGTTGAAGATGCTTCTATTGCTGATAAAGCTTTTTCACTGTTGATGGGTGATGAAGTTCCCCCAAGAAGAGAATTTATTGAAAAAAATGCTCGTAGTGTTAAAAACATTGATATTTAA
>Myco_7294_8547_+
ATGTTGGATCCAAACAAATTACGCAATAACTATGATTTCTTTAAAAAGAAACTGTTAGAAAGAAATGTAAATGAGCAATTATTAAATCAGTTTATTCAAACTGATAAACTAATGCGCAAAAACTTGCAACAACTTGAACTTGCTAACCAAAAACAAAGCTTGTTGGCAAAACAAGTTGCTAAGCAAAAAGATAATAAAAAGCTATTAGCTGAATCAAAAGAACTTAAGCAGAAGATTGAAAACTTAAATAATGCTTATAAAGATTCACAAAACATTAGTCAAGATTTACTTCTAAATTTTCCTAATATTGCTCATGAATCAGTTCCTGTTGGTAAAAATGAATCAGCAAACTTAGAACTTCTTAAAGAAGGGAGAAAACCAGTTTTTGATTTCAAACCTTTACCACATCGAGAGTTATGTGAAAAGTTAAATTTAGTTGCTTTTGATAAAGCTACTAAGATTAGTGGAACTAGGTTTGTTGCATATACAGATAAAGCAGCTAAACTACTTAGAGCGATAACTAATCTAATGATTGACCTTAATAAAAGCAAGTATCAAGAATGAAACCTGCCAGTTGTTATTAATGAATTAAGTTTAAGATCAACCGGACAACTACCTAAGTTTAAAGATGATGTTTTTAAACTAGAAAACACCCGTTATTATCTTTCTCCAACTTTAGAGGTACAACTTATCAATTTACATGCTAATGAAATTTTTAATGAAGAAGATTTACCTAAATACTACACTGCAACAGGTATTAACTTTCGTCAAGAAGCGGGTAGTGCTGGTAAACAAACCAAAGGAACTATTAGATTGCATCAGTTTCAAAAAACTGAGTTAGTTAAGTTTTGTAAACCTGAAAATGCTATCAATGAATTGGAAGCAATGGTTAGAGATGCTGAACAAATCTTAAAGGCACTTAAGTTACCTTTTAGAAGGTTATTGTTATGTACTGGTGATATGGGCTTTAGTGCTGAAAAAACATATGATCTTGAAGTTTGAATGGCAGCTAGCAATGAATATCGTGAAGTTTCTTCTTGTTCATCTTGTGGTGATTTTCAAGCAAGAAGAGCTATGATTCGTTACAAAGATATTAACAACGGTAAAAACAGTTATGTTGCTACTTTAAATGGAACAGCATTATCTATTGATAGAATTTTTGCTGCAATTCTAGAAAATTTTCAAACAAAAGATGGCAAAATTCTTATCCCACAAGCATTAAAAAAATACCTTGATTTTGACACAATCAAGTAA
......
ORFs_Without_Corresponding_Gene_In_Reference_Metrics:
ATG_Start ,GTG_Start ,TTG_Start ,ATT_Start ,CTG_Start ,Alternative_Start_Codon ,TGA_Stop ,TAA_Stop ,TAG_Stop ,Alternative_Stop_Codon ,Median_Length ,ORFs_on_Positive_Strand ,ORFs_on_Negative_Strand
58.39,17.14,24.47,0.00,0.00,0.00,71.55,20.62,7.83,0.00,287.00,449,356
ORF_Without_Corresponding_Gene_in_Reference:
>Prodigal_1828_2073_+
ATGAATCTTTACGATCTTTTAGAACTACCAACTACAGCATCAATAAAAGAAATAAAAATTGCTTATAAAAGATTAGCAAAGCGTTATCACCCTGATGTAAATAAATTAGGTTCGCAAACTTTTGTTGAAATTAATAATGCTTATTCAATATTAAGTGATCCTAACCAAAAGGAAAAATATGATTCAATGCTGAAAGTTAATGATTTTCAAAATCGCATCAAAAATTTAGATATTAGTGTTAGATGA
>Prodigal_2605_2760_+
ATGAAAGTAGTTAATAAAGTAAACAAAAGACTGCGTATTTTTTCAAGCTTTTTTGAGAACGATAAATCTAAATTATGGTTCCTTGTTCCAAACGATAAACAAAGTAATCCTAATAAGGGCGTTTTTAACTATAAAACTCAGCACTTTATTGATTAA
>Prodigal_2845_2979_+
ATGGAAGAAAATAACAAAGCAAATATCTATGACTCTAGTAGCATTAAGGTCCTTGAAGGACTTGAGGCTGTTAGAAAACGCCCTGGAATGTACATTGGTTCTACTGGCGAAGAAGGTTTGCATCACATGATCTGA
>Prodigal_3010_3255_+
ATGGGAGGTTTTGCCAGTTTTGTTAAGCTTACCCTTGAAGATAATTTTGTTACCCGTGTAGAGGATGATGGAAGAGGGATACCTGTTGATATCCATCCTAAGACTAATCGTTCTACAGTTGAAACAGTTTTTACAGTTCTACACGCTGGCGGTAAATTTGATAACGATAGCTATAAAGTGTCAGGTGGTTTACACGGTGTTGGTGCATCAGTTGTTAATGCGCTTAGTTCTTCTTTTAAAGTTTGA
>Prodigal_3319_3513_+
TTGGTCCAAGAAGGTAACTCTGAAAAAGAGCATGGAACAATTGTTGAGTTTGTTCCTGATTTCTCTGTAATGGAAAAGAGTGATTACAAACAAACTGTAATTGTAAGCAGACTCCAGCAATTAGCTTTTTTAAACAAGGGAATAAGAATTGACTTTGTTGATAATCGTAAACAAAACCCACAGTCTTTTTCTTGA
>Prodigal_3529_4557_+
TTGGTTGAATATATCCACCACCTAAACAACGAAAAAGAACCACTTTTTAATGAAGTTATTGCTGATGAAAAAACTGAAACTGTAAAAGCTGTTAATCGTGATGAAAACTACACAGTAAAGGTTGAAGTTGCTTTTCAATATAACAAAACATACAACCAATCAATTTTCAGTTTTTGTAACAACATTAATACTACAGAAGGTGGAACCCATGTGGAAGGTTTTCGTAATGCACTTGTTAAGATCATTAATCGCTTTGCTGTTGAAAATAAATTCCTAAAAGATAGTGATGAAAAGATTAACCGTGATGATGTTTGTGAAGGATTAACTGCTATTATTTCCATTAAACACCCAAACCCACAATATGAAGGACAAACTAAAAAGAAGTTAGGTAATACTGAGGTAAGACCTTTAGTTAATAGTGTTGTTAGTGAAATCTTTGAACGCTTCATGTTAGAAAACCCACAAGAAGCAAACGCTATCATCAGAAAAACACTTTTAGCTCAAGAAGCGAGAAGAAGAAGTCAAGAGGCTAGGGAGTTAACTCGTCGTAAATCACCTTTTGATAGTGGTTCATTACCAGGTAAATTAGCTGATTGTACAACCAGAGATCCTTCGATTAGTGAACTTTACATTGTTGAGGGTGATAGTGCTGGTGGCACTGCTAAAACAGGAAGAGATCGTTATTTTCAAGCTATCTTACCCTTAAGAGGAAAGATTTTAAACGTTGAAAAATCTAACTTTGAACAAATCTTTAATAATGCAGAAATTTCTGCATTAGTGATGGCAATAGGCTGTGGGATTAAACCTGATTTTGAACTTGAAAAACTTAGATATAGCAAGATTGTGATCATGACAGATGCTGATGTTGATGGTGCACACATAAGAACACTTCTCTTAACTTTCTTTTTTCGCTTTATGTATCCTTTGGTTGAACAAGGCAATATTTTTATTGCTCAACCCCCACTTTATAAAGTGTCATATTCCCATAAGGATTTATACATGCACACTGATGTTCAACTTGAACAGTGA
....
ORFs_Which_Detected_more_than_one_Gene:
GFF Tools:
GFF-Adder:
GFF-Adder allows for the addition of predicted CDSs to an existing reference annotation (GFF or another tool) which produces a new GFF containing the original
genes plus the new CDS from another prediction. Default filtering will remove additional CDSs that overlap existing genes by more than 50 nt.
The -gi
option can be used to allow for different genomic elements to be accounted for, other than only CDSs in the reference annotation.
For Help: GFF-Adder -h
Thank you for using ORForise
Please report any issues to: https://github.com/NickJD/ORForise/issues
#####
usage: GFF_Adder.py [-h] -dna GENOME_DNA -ref REFERENCE_ANNOTATION -at ADDITIONAL_TOOL -add ADDITIONAL_ANNOTATION -o
OUTPUT_FILE [-rt REFERENCE_TOOL] [-gi GENE_IDENT] [-gene_ident GENE_IDENT] [-olap OVERLAP]
ORForise v1.4.2: GFF-Adder Run Parameters.
Required Arguments:
-dna GENOME_DNA Genome DNA file (.fa) which both annotations are based on
-ref REFERENCE_ANNOTATION
Which reference annotation file to use as reference?
-at ADDITIONAL_TOOL Which format to use for additional annotation?
-add ADDITIONAL_ANNOTATION
Which annotation file to add to reference annotation?
-o OUTPUT_FILE Output filename
Optional Arguments:
-rt REFERENCE_TOOL Which tool format to use as reference? - If not provided, will default to standard Ensembl
GFF format, can be Prodigal or any of the other tools available
-gi GENE_IDENT Identifier used for extraction of "genic" regions from reference annotation "CDS,rRNA,tRNA":
Default for is "CDS"
-gene_ident GENE_IDENT
Identifier used for identifying genomic features in reference annotation "CDS,rRNA,tRNA"
-olap OVERLAP Maximum overlap between reference and additional genic regions (CDS,rRNA etc) - Default: 50
nt
Example: Running GFF-Adder to combine the additional CDS predictions made by Prodial to the canonical annotations from Ensembl.
GFF-Adder -dna ~/Testing/Myco.fa -ref ~/Testing/Myco.gff -at Prodigal -add ~/Testing/Prodigal_Myco.gff -o ~/Testing/Myco_Ensembl_GFF_Adder_Prodigal.gff
Example Output: ~/ORForise/Testing/Myco_Ensembl_GFF_Adder_Prodigal.gff
##gff-version 3
# GFF-Adder
# Run Date:2021-11-10
##Genome DNA File:./Testing/Myco.fa
##Original File: ./Testing/Myco.gff
##Additional File: ./Testing/Prodigal_Myco.gff
.......
Chromosome Reference_Annotation CDS 68522 70225 . - . ID=Original_Annotation
Chromosome Reference_Annotation CDS 70530 72572 . + . ID=Original_Annotation
Chromosome Reference_Annotation CDS 72523 73434 . + . ID=Original_Annotation
Chromosome Prodigal CDS 73445 73648 . + . ID=Additional_Annotation
Chromosome Reference_Annotation CDS 73690 77685 . + . ID=Original_Annotation
Chromosome Reference_Annotation CDS 77685 79085 . + . ID=Original_Annotation
Chromosome Reference_Annotation CDS 79089 81035 . + . ID=Original_Annotation
Chromosome Reference_Annotation CDS 81046 82596 . + . ID=Original_Annotation
Chromosome Reference_Annotation CDS 82620 84044 . + . ID=Original_Annotation
Chromosome Prodigal CDS 84082 84312 . + . ID=Additional_Annotation
Chromosome Prodigal CDS 84532 84744 . - . ID=Additional_Annotation
Chromosome Prodigal CDS 84776 85051 . + . ID=Additional_Annotation
GFF-Intersector:
GFF-Intersector enables the aggregation of different genome annotations and CDS predictions and creates a single GFF
representing the intersection of the two existing annotations.
GFF-Intersector also provides an option to allow the retention of genes that have a user defined difference (minimum % coverage and in-frame).
The -gi
option can be used to allow for different genomic elements to be accounted for, other than only CDSs in the reference annotation.
For Help: GFF-Intersector -h
Thank you for using ORForise
Please report any issues to: https://github.com/NickJD/ORForise/issues
#####
usage: GFF_Intersector.py [-h] -dna GENOME_DNA -ref REFERENCE_ANNOTATION -at ADDITIONAL_TOOL -add
ADDITIONAL_ANNOTATION -o OUTPUT_FILE [-rt REFERENCE_TOOL] [-gi GENE_IDENT] [-cov COVERAGE]
ORForise v1.4.2: GFF-Intersector Run Parameters.
Required Arguments:
-dna GENOME_DNA Genome DNA file (.fa) which both annotations are based on
-ref REFERENCE_ANNOTATION
Which reference annotation file to use as reference?
-at ADDITIONAL_TOOL Which format to use for additional annotation?
-add ADDITIONAL_ANNOTATION
Which annotation file to add to reference annotation?
-o OUTPUT_FILE Output filename
Optional Arguments:
-rt REFERENCE_TOOL Which tool format to use as reference? - If not provided, will default to standard Ensembl
GFF format, can be Prodigal or any of the other tools available
-gi GENE_IDENT Identifier used for extraction of "genic" regions from reference annotation "CDS,rRNA,tRNA":
Default for is "CDS"
-cov COVERAGE Percentage coverage of reference annotation needed to confirm intersection - Default: 100 ==
exact match
Example: Running GFF-Intersector to combine the additional CDS predictions made by Prodial to the canonical annotations from Ensembl.
GFF-Intersector -dna ~/Testing/Myco.fa -ref ~/Testing/Myco.gff -at Prodigal -add ~/Testing/Prodigal_Myco.gff -o ~/Testing/Myco_Ensembl_GFF_Intersector_Prodigal.gff
Example Output: ~/Testing/Myco_Ensembl_GFF_Intersector_Prodigal.gff
##gff-version 3
# GFF-Intersector
# Run Date:2021-11-10
##Genome DNA File:./Testing/Myco.fa
##Original File: ./Testing/Myco.gff
##Intersecting File: ./Testing/Prodigal_Myco.gff
Chromosome original CDS 686 1828 . + . ID=Original_Annotation;Coverage=100
Chromosome original CDS 4812 7322 . + . ID=Original_Annotation;Coverage=100
Chromosome original CDS 8551 9183 . + . ID=Original_Annotation;Coverage=100
Chromosome original CDS 22389 23558 . + . ID=Original_Annotation;Coverage=100
Chromosome original CDS 29552 30124 . + . ID=Original_Annotation;Coverage=100
Chromosome original CDS 31705 32325 . - . ID=Original_Annotation;Coverage=100
Chromosome original CDS 49376 49642 . + . ID=Original_Annotation;Coverage=100
Chromosome original CDS 59082 59753 . + . ID=Original_Annotation;Coverage=100
Chromosome original CDS 61014 61406 . + . ID=Original_Annotation;Coverage=100
Chromosome original CDS 82620 84044 . + . ID=Original_Annotation;Coverage=100
Genomes Available:
The .fa and .gff files (from Ensembl Bacteria Release 46) below are available in the Genomes directory.
- Bacillus subtilis - Strain BEST7003 - Assembly ASM52304v1
- Caulobacter crescentus - Strain CB15 - Assembly ASM690v1
- Escherichia coli K-12 - Strain ER3413 - Assembly ASM80076v1
- Mycoplasma genitalium - Strain G37 - Assembly ASM2732v1
- Pseudomonas fluorescens - Strain UK4 - Assembly ASM73042v1
- Staphylococcus aureus - Strain 502A - Assembly ASM59796v1
Prediction Tools Available:
There are two Groups of tools - Those which do require a pre-built model and those which do not.
For the example runs provided, each tool is listed with the non-default options used and their predictions for each of the 6 model organisms are available in their respective
directories.
ORForise only needs the tool name and the annotation file produced from any available model to undertake the analysis.
GFF Standard Format:
The GFF Tool directory allows for the analysis of user-provided annotations in the standard GFF3 format.
This can be used to compare different cannonical annotations with eachother or additional tools which use the GFF3 format.
Model Based Tools:
Augustus - Version 3.3.3 - http://bioinf.uni-greifswald.de/augustus/
This tool has three comparisons with the organism models E. coli and S. aureus and H. sapiens.
EasyGene - Version 1.2 - http://www.cbs.dtu.dk/services/EasyGene/
This tool has two comparisons with the organism models E. coli - K12 and S. aureus Mu50.
FGENESB - Version '2020' - http://www.softberry.com/berry.phtml?topic=fgenesb&group=programs&subgroup=gfindb
This tool has two comparisons with the organism models E. coli - K12 and S. aureus MU50.
GeneMark - Version 2.5 - http://exon.gatech.edu/GeneMark/gm.cgi
This tool has two comparisons with the organism models E. coli - K12 - MG165 and S. aureus Mu50.
GeneMark.hmm - Version 3.2.5 - http://exon.gatech.edu/GeneMark/gmhmmp.cgi
This tool has two comparisons with the organism models E. coli - K12 - MG165 and S. aureus Mu50.
Self-Training/Non-Model Based Tools
FragGeneScan - Version 1.3.0 - https://omics.informatics.indiana.edu/FragGeneScan/
The 'complete' genome option was selected and GFF was chosen as output type.
GeneMark HA - Version 3.25 - http://exon.gatech.edu/GeneMark/heuristic_gmhmmp.cgi
GFF was chosen as output type.
GeneMarkS - Version 4.25 - http://exon.gatech.edu/GeneMark/genemarks.cgi
GFF was chosen as output type.
GeneMarkS-2 - Version '2020' - http://exon.gatech.edu/GeneMark/genemarks2.cgi
GFF3 was chosen as output type.
GLIMMER-3 - Version 3.02 - http://ccb.jhu.edu/software/glimmer/index.shtml
Default parameters from manual were used.
MetaGene - Version 2.24.0 - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1636498/
Default options were used.
MetaGeneAnnotator - Version '2008/8/19' - http://metagene.nig.ac.jp/
Defaults options were used.
MetaGeneMark - Version '2020' - http://exon.gatech.edu/meta_gmhmmp.cgi
GFF was chosen as output type.
Prodigal - Version 2.6.3 - https://github.com/hyattpd/Prodigal
GFF was chosen as output type.
TransDecoder - Version 5.5.0 - https://github.com/TransDecoder/TransDecoder/wiki
Defaults options were used.
Balrog - Version 2021` - https://github.com/salzberg-lab/Balrog
Defaults options were used.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.