[maker-devel] updating evidence based gff with IPRScan results

[Dinatale C]

Hi Carson,

  A little while back you provided me with a script to edit my gff files in order to update them with IPRscan resutls. It seems I am able to do so when only maker models are present in the gff file, however, when I try to update an anlysis/evidence based gff file I recieve the errors below. I have attached the files for you to look at, any help is appreciated. (  i attached all sets of files, those that worked, 'models', and those that did not, 'abinit')

Thanks in advance,
Claudia

---

ERROR ( is associated with all lines in the IPRSCAN file)

Use of uninitialized value $gene_id in hash element at /usr/local/genome/maker/maker-2.10/bin/ipr_update_gff line 163, <$IN> line 61.

[Carson Holt]

The script is working with the gene attributes so results where you have interpro domains for match/match_part entries causes the error.  You can use the gff3_preds2models script to convert the set of ab initio prediction entries into gene/mRNA/exon/CDS models.  You just provide a list of what predictions you want to convert (1 ID per line).

--Carson

[claudia]

Hi,
 I was wondering if anyone can tell me exactly how the maker_functional_gff script works?

 I have attempted to update a gff3 with a  Wu-blast output of maker proteins ( and specifying uniprot-sprot db), as well, I have downloaded the uniprot sprot fasta of the most up to date protein db in fasta format ( not sure if this is the right file to use) and this is what i used as the 'uniprot-sprot fasta', but Im not sure if this is correct.

Basically, the gff3 note attribute gets updated with only "protein of unknown function", yet I know that there are specific protein hits that it is not utilizing, as well these protein hits are located in both the wu-blast output and the uniprot-sprot db fasta file.

Thanks,
Claudia

[Carson Holt]

Usage:

maker_functional_gff uniprot-sprot.fasta uniprot-sprot.blast.out maker.gff

The uniprot-sprot.fasta file is just the raw uniprot fasta file

The uniprot-sprot.blast.out file is the WUBLAST results using –mformat=2 option (tab delimited).  These are blast results for the MAKER proteins against the uniprot/swiss-prot dataset. Ie -->

blastp  uniprot-sprot.fasta  maker_proteins.fasta -mformat-=2 > uniprot-sprot.blast.out

(order of files is important maker_proteins must come second)

maker.gff is just a MAKER produced GFF3 file.


The script takes the best hit for each maker protein and keeps that function.

--Carson

Carson Holt
Graduate Student
Yandell Lab
http:/www.yandell-lab.org/
Eccles Institute of Human Genetics
University of Utah

[claudia]

Hi Carson,
 Thanks for the reply, I see how it works now, I was generating a wu-blast output from the actual website.
One thing, if I am using NCBI blast, will this script still work? I think the formating is different from wu-blast.

Thanks, Claudia

[Carson Holt]

You can try NCBI’s tab delimited output format.  You may have to change the order of columns.  First column must be the query, second the hit, and third the score.  I believe all other columns are ignored.

--Carson

On 5/11/11 12:43 PM, "claudia" <din...@uwindsor.ca> wrote:

 Hi Carson,
  Thanks for the reply, I see how it works now, I was generating a wu-blast output from the actual website.
 One thing, if I am using NCBI blast, will this script still work? I think the formating is different from wu-blast.
 
 Thanks, Claudia
 
 
 On 11/05/2011 2:23 PM, Carson Holt wrote:

Re: updating gff3 note attribute with wu-blast hits using maker_functional_gff Usage:

[Claudia DiNatale]

Hi Carson,
 I have changed the column positions as you mentioned, however I get the error stated below, I have attached the 2 files I am using, and  I am using  the uniprot_sprot.fasta downloaded from the uniprot site .



Thanks, Claudia


##gff-version 3
Can't use string ("") as a HASH ref while "strict refs" in use at /usr/local/genome/maker/maker-2.10/bin/maker_functional_gff line 58, <$IN> line 2.

[Carson Holt]

I’m not sure how you’re running this.  Your results show a hit named  “gi|19924280|sp|P49693.3|RL193_ARATH”.

Uniprot/Swiss-prot fasta headers from the EBI look like this -->

>sp|Q76N59|ZYG1_DICMU Zygote formation protein zyg1 OS=Dictyostelium mucoroides GN=zyg1 PE=2 SV=1

The script is expecting a very specific header format that would probably not be found in GenBank (I can tell from gi identifier in your BLAST hit names).

Correct file-->
ftp://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.fasta.gz

You will need to run the blast locally.

Thanks,
Carson

[Carson Holt]

When you run blast locally, you need to create the index yourself.

NCBI BLAST use formatdb
NCBI BLAST+ used makeblastdb

I.e -->

[cholt@garrucha blastdb]$  /usr/local/ncbi-blast-2.2.23+/bin/makeblastdb -in uniprot_sprot.fasta

Building a new DB, current time: 05/11/2011 16:32:20
New DB name:   uniprot_sprot.fasta
New DB title:  uniprot_sprot.fasta
Sequence type: Protein
Keep Linkouts: T
Keep MBits: T
Maximum file size: 1073741824B
Adding sequences from FASTA; added 495880 sequences in 27.28 seconds

Thanks,
Carson

On 5/11/11 4:29 PM, "Claudia DiNatale" <claudia...@hotmail.com> wrote:

 Hi,
  So I am running the NCBI blastp locally, and the script  asks you to specify the database on the command line, when I tried to specify using the uniprot-swissprot fasta file, it returns an error saying no alias or index id specified. To fix this, I specifically downloaded the swissprot db from the ftp site at NCBI (which has a particular index format that the script understands) and this db requires the parent database ( non-redundant proteins) to also be present, so basically the end result is that I get both the NCBI and swiss prot identifiers in the blast output. I could not run blastp with just swissprot as the db, which I assume would have given me just the sp identifier. So I am not sure how to get around this.
 
  And, yes I am using that uniprot_sprot fasta file you have attached.
 
 Claudia
 
 
 On 11/05/2011 6:19 PM, Carson Holt wrote:

Re: updating gff3 note attribute with wu-blast hits using maker_functional_gff I’m not sure how you’re running this.  Your results show a hit named  “gi|19924280|sp|P49693.3|RL193_ARATH”.

[Claudia DiNatale]

Hi,
 So I am running the NCBI blastp locally, and the script  asks you to specify the database on the command line, when I tried to specify using the uniprot-swissprot fasta file, it returns an error saying no alias or index id specified. To fix this, I specifically downloaded the swissprot db from the ftp site at NCBI (which has a particular index format that the script understands) and this db requires the parent database ( non-redundant proteins) to also be present, so basically the end result is that I get both the NCBI and swiss prot identifiers in the blast output. I could not run blastp with just swissprot as the db, which I assume would have given me just the sp identifier. So I am not sure how to get around this.

 And, yes I am using that uniprot_sprot fasta file you have attached.

Claudia


On 11/05/2011 6:19 PM, Carson Holt wrote:

[Dinatale C]

Hi,

  So I have attempted to use the add_utr_to_gff3 script and it seems to return many errors, see below: ( I have attached the gff3 file I was trying to update)

---

Global symbol "$s" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 26.
Global symbol "@files" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 34.
Global symbol "@files" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 40.
Global symbol "@files" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 48.
Global symbol "@stuff" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 63.
Global symbol "@stuff" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 86.
Global symbol "@stuff" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 87.
Global symbol "@stuff" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 88.
Global symbol "$header" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 93.
Global symbol "$footer" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 96.
Global symbol "@stuff" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 111.
Global symbol "$footer" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 187.
Global symbol "@regions" requires explicit package name at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 445.
Missing right curly or square bracket at /usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 line 489, at end of line
/usr/local/genome/maker/maker-2.10/bin/add_utr_to_gff3 has too many errors.

---

so I then tried the other script add_utr_start_stop_codon_gff instead and the script added the features however it seems to change all of my I.D's to an increasing numerical order, see below:

---

##gff-version 3
contig00005    maker    gene    78618    82008    .    +    .    ID=1;Name=augustus_masked-contig00005-abinit-gene-0.15
contig00005    maker    mRNA    78618    82008    .    +    .    ID=2;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1;Parent=1
contig00005    maker    exon    78618    78709    0.04    +    .    ID=3;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:exon:0;Parent=2
contig00005    maker    exon    79771    79837    0.04    +    .    ID=4;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:exon:1;Parent=2
contig00005    maker    exon    80234    81429    0.04    +    .    ID=5;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:exon:2;Parent=2
contig00005    maker    exon    81587    82008    6726    +    .    ID=6;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:exon:3;Parent=2
contig00005    maker    five_prime_UTR    78618    78709    .    +    .    ID=7;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:UTR1;Parent=2
contig00005    maker    five_prime_UTR    79771    79837    .    +    .    ID=8;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:UTR2;Parent=2
contig00005    maker    five_prime_UTR    80234    80556    .    +    .    ID=9;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:UTR3;Parent=2
contig00005    maker    CDS    80557    81429    .    +    0    ID=10;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:cds:0;Parent=2
contig00005    maker    CDS    81587    81910    .    +    0    ID=11;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:cds:1;Parent=2
contig00005    maker    three_prime_UTR    81911    82008    .    +    .    ID=12;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:UTR4;Parent=2
contig00005    maker    start_codon    80557    80559    .    +    .    ID=13;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:start1;Parent=2
contig00005    maker    stop_codon    81908    81910    .    +    .    ID=14;Name=augustus_masked-contig00005-abinit-gene-0.15-mRNA-1:stop2;Parent=2

---

Does anyone have any idea how to control this so that my I.D's don't change, and/ or how to use the other script without getting the errors?

Thanks in advance,

Claudia

[Carson Holt]

The add_utr_to_gff3 script should have been pulled out, and somehow was left in the 2.10 release.  The add_utr_start_stop_codon_gff script changes IDs, which normally isn’t too big a problem because chado will also change them (ID’s just need to be unique and are not necessarily informative).

If you need those to be unchanged, I’ll modify the script.  It will be a few days though because I’m defending my thesis on Monday.

You can also use MAKER 2.11-beta.  Named UTR is supposed to be optional in GFF3 format because it is always implicit anyway, but MAKER 2.11 now adds explicit UTR by default to better work with JBrowse.  Just give it the other MAKER’s GFF3 file and turn all the pass options to 1.  Don’t use other input files including repeatmasker options.

Thanks,
Carson

[Khan, Anar]

Hi Carson

 

I’m running maker version 2.09 and would like to add UTRs to the gff output. I’m getting errors for both the add_utr_start_stop_gff and add_utr_to_gff3 scripts.

 

I see from your previous post that an upgrade is recommended but I don’t think our license includes upgrades.

 

I was just wondering if you have any ideas on how to obtain UTR info using v2.09? The script errors are pasted below.

 

Cheers

Anar

 

add_utr_to_gff3 test_maker.orig.gff

Global symbol "$s" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 26.

Global symbol "@files" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 34.

Global symbol "@files" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 40.

Global symbol "@files" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 48.

Global symbol "@stuff" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 63.

Global symbol "@stuff" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 86.

Global symbol "@stuff" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 87.

Global symbol "@stuff" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 88.

Global symbol "$header" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 93.

Global symbol "$footer" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 96.

Global symbol "@stuff" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 111.

Global symbol "$footer" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 187.

Global symbol "@regions" requires explicit package name at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 445.

Missing right curly or square bracket at /usr/local/maker_2.09beta/bin/add_utr_to_gff3 line 489, at end of line

/usr/local/maker_2.09beta/bin/add_utr_to_gff3 has too many errors.

 

--------------

 

add_utr_start_stop_gff test_maker.orig.gff

Use of uninitialized value in split at /usr/local/maker_2.09beta/bin/add_utr_start_stop_gff line 89.

Use of uninitialized value in substr at /usr/local/maker_2.09beta/bin/add_utr_start_stop_gff line 305.

substr outside of string at /usr/local/maker_2.09beta/bin/add_utr_start_stop_gff line 305.

Use of uninitialized value in substr at /usr/local/maker_2.09beta/bin/add_utr_start_stop_gff line 306.

substr outside of string at /usr/local/maker_2.09beta/bin/add_utr_start_stop_gff line 306.

Use of uninitialized value in print at /usr/local/maker_2.09beta/bin/add_utr_start_stop_gff line 180.

 

---

Attention: The information contained in this message and/or attachments from AgResearch Limited is intended only for the persons or entities to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipients is prohibited by AgResearch Limited. If you have received this message in error, please notify the sender immediately.

---

 

[Barry Moore]

Hi Anar,

I looks like the add_utr_to_gff3 script is completely hosed in that release.  I think it's just there as a relic and add_utr_start_stop_gff is the script that you want.  From those errors it looks like your gff3 file doesn't have the fasta sequence at the end of it. 

If I've guessed correctly, the quick workaround would be to just add lines like the following to the end of your GFF3 file - where ,of course, the fasta header and sequence are replaced with the correct data for the chromosome(s)/contig(s) represented in your GFF3 file. 

##FASTA

>chr1

ATCGATACGTAG...

B

<ATT00001..txt>

Barry Moore

Research Scientist

Dept. of Human Genetics

University of Utah

Salt Lake City, UT 84112

--------------------------------------------

(801) 585-3543

[Carson Holt]

2.11 and above no longer have those scripts (UTR is always added by default now).  So this is a round about way to do it --> upgrade MAKER, then provide the merged GFF3 (made with gff3_merge) to genome_gff in the makeropts.ctl file, then set all pass options to 1.  Then make sure to clear all other options (repeatprotein model_org etc.).  When you rerun MAKER, it will just update the GFF3.

--Carson

[Khan, Anar]

Hi

 

Thank you both for your replies.

 

Barry, you reminded me that I had in fact meddled with the gff output, so what I was passing to the add_utr scripts wasn’t actually bona fide MAKER output. I had stripped the fasta from the gff file, for some reason. Running add_utr_start_stop_gff on the original MAKER gff worked without errors. My mistake, I apologise! Thank you for picking this up.

 

Carson, thanks for explaining the workaround. We can’t afford to pay for an upgrade right now, but will obtain the new version when our current license expires and we renew. I’m happy to hear MAKER now adds UTRs by default.

 

Cheers

Anar

[Khan, Anar]

Hi

 

I apologise for asking another question about this script given it’s been deprecated, but I am curious about something and wonder whether you might be able to help. I ran add_utr_start_stop_gff on some maker output and no UTRs were added. For this species I have no est data, only altest data. Does this explain why no UTRs could be identified? What info is used to add UTR coordinates?

 

Thanks very much,

Anar

 

 

From: Carson Holt [mailto:carso...@genetics.utah.edu]

Sent: Wednesday, 14 September 2011 2:02 p.m.
To: Barry Moore; Khan, Anar

Cc: Dinatale C; maker...@yandell-lab.org
Subject: Re: [maker-devel] usage of add_utr scripts

 

2.11 and above no longer have those scripts (UTR is always added by default now).  So this is a round about way to do it --> upgrade MAKER, then provide the merged GFF3 (made with gff3_merge) to genome_gff in the makeropts.ctl file, then set all pass options to 1.  Then make sure to clear all other options (repeatprotein model_org etc.).  When you rerun MAKER, it will just update the GFF3.

 

--Carson

 

 

 

From: Barry Moore <bmo...@genetics.utah.edu>
Date: Tue, 13 Sep 2011 18:41:16 -0600
To: "Khan, Anar" <Anar...@agresearch.co.nz>
Cc: Carson Holt <carso...@genetics.utah.edu>, Dinatale C <din...@uwindsor.ca>, "maker...@yandell-lab.org" <maker...@yandell-lab.org>
Subject: Re: [maker-devel] usage of add_utr scripts

 

Hi Anar,

 

I looks like the add_utr_to_gff3 script is completely hosed in that release.  I think it's just there as a relic and add_utr_start_stop_gff is the script that you want.  From those errors it looks like your gff3 file doesn't have the fasta sequence at the end of it. 

 

If I've guessed correctly, the quick workaround would be to just add lines like the following to the end of your GFF3 file - where ,of course, the fasta header and sequence are replaced with the correct data for the chromosome(s)/contig(s) represented in your GFF3 file. 

 

##FASTA

>chr1

ATCGATACGTAG...

 

B

 

[Carson Holt]

ESTs that overlap the splice sites of the prediction are used to extend the gene model.  You can think of it as a kind of cut and paste operation.  Alt_est does not match the splice sites correctly (there are several technical reasons why).  Also since  alt_est is aligned in protein space (tblastx) the UTR region may not line up as it is non-coding, so it is not under the same evolutionary constraints as the rest of the alignment.

Thanks,

Carson

From: "Khan, Anar" <Anar...@agresearch.co.nz>
Date: Wed, 28 Sep 2011 19:07:53 -0600
To: Carson Holt <carso...@genetics.utah.edu>, Barry Moore <bmo...@genetics.utah.edu>
Cc: Dinatale C <din...@uwindsor.ca>, "maker...@yandell-lab.org" <maker...@yandell-lab.org>
Subject: RE: [maker-devel] usage of add_utr scripts

Hi

 

I apologise for asking another question about this script given it’s been deprecated, but I am curious about something and wonder whether you might be able to help. I ran add_utr_start_stop_gff on some maker output and no UTRs were added. For this species I have no est data, only altest data. Does this explain why no UTRs could be identified? What info is used to add UTR coordinates?

 

Thanks very much,

Anar

 

 

From: Carson Holt [mailto:carso...@genetics.utah.edu]
Sent: Wednesday, 14 September 2011 2:02 p.m.
To: Barry Moore; Khan, Anar
Cc: Dinatale C; maker...@yandell-lab.org
Subject: Re: [maker-devel] usage of add_utr scripts

 

2.11 and above no longer have those scripts (UTR is always added by default now).  So this is a round about way to do it --> upgradeMAKER, then provide the merged GFF3 (made with gff3_merge) to genome_gff in the makeropts.ctl file, then set all pass options to 1.  Then make sure to clear all other options (repeatprotein model_org etc.).  When you rerun MAKER, it will just update the GFF3.

 

--Carson

 

 

 

From: Barry Moore <bmo...@genetics.utah.edu>
Date: Tue, 13 Sep 2011 18:41:16 -0600
To: "Khan, Anar" <Anar...@agresearch.co.nz>
Cc: Carson Holt <carso...@genetics.utah.edu>, Dinatale C <din...@uwindsor.ca>, "maker...@yandell-lab.org" <maker...@yandell-lab.org>
Subject: Re: [maker-devel] usage of add_utr scripts

 

Hi Anar,

 

I looks like the add_utr_to_gff3 script is completely hosed in that release.  I think it's just there as a relic and add_utr_start_stop_gff is the script that you want.  From those errors it looks like your gff3 file doesn't have the fasta sequence at the end of it. 

 

If I've guessed correctly, the quick workaround would be to just add lines like the following to the end of your GFF3 file - where ,of course, the fasta header and sequence are replaced with the correct data for the chromosome(s)/contig(s) represented in your GFF3 file. 

 

##FASTA

>chr1

ATCGATACGTAG...

 

B

 

[Khan, Anar]

Hi Carson

 

Thanks a lot for the explanation, all clear now.

 

Cheers

[Carson Holt]

Yes.  That would be why.

--Carson

From: "Khan, Anar" <Anar...@agresearch.co.nz>
Date: Wed, 28 Sep 2011 18:59:04 -0600
To: Carson Holt <carso...@genetics.utah.edu>, Barry Moore <bmo...@genetics.utah.edu>
Cc: Dinatale C <din...@uwindsor.ca>, "maker...@yandell-lab.org" <maker...@yandell-lab.org>
Subject: RE: [maker-devel] usage of add_utr scripts

Hi

 

I ran add_utr_start_stop_gff on some maker output and no UTRs were added. For this species I have no est data, only altest data. Does this explain why no UTRs could be identified –perhaps altest evidence isn’t used when attempting to define UTRs?

 

Cheers :-)

Anar

 

From: Khan, Anar
Sent: Wednesday, 14 September 2011 4:19 p.m.
To: 'Carson Holt'; Barry Moore
Cc: Dinatale C; maker...@yandell-lab.org
Subject: RE: [maker-devel] usage of add_utr scripts

 

Hi

 

Thank you both for your replies.

 

Barry, you reminded me that I had in fact meddled with the gff output, so what I was passing to the add_utr scripts wasn’t actually bona fide MAKER output. I had stripped the fasta from the gff file, for some reason. Running add_utr_start_stop_gff on the original MAKER gff worked without errors. My mistake, I apologise! Thank you for picking this up.

 

Carson, thanks for explaining the workaround. We can’t afford to pay for an upgrade right now, but will obtain the new version when our current license expires and we renew. I’m happy to hear MAKER nowadds UTRs by default.

 

Cheers

Anar

 

 

From: Carson Holt [mailto:carso...@genetics.utah.edu]
Sent: Wednesday, 14 September 2011 2:02 p.m.
To: Barry Moore; Khan, Anar
Cc: Dinatale C; maker...@yandell-lab.org
Subject: Re: [maker-devel] usage of add_utr scripts

 

2.11 and above no longer have those scripts (UTR is always added by default now).  So this is a round about way to do it --> upgradeMAKER, then provide the merged GFF3 (made with gff3_merge) to genome_gff in the makeropts.ctl file, then set all pass options to 1.  Then make sure to clear all other options (repeatprotein model_org etc.).  When you rerun MAKER, it will just update the GFF3.

 

--Carson

 

 

 

From: Barry Moore <bmo...@genetics.utah.edu>
Date: Tue, 13 Sep 2011 18:41:16 -0600
To: "Khan, Anar" <Anar...@agresearch.co.nz>
Cc: Carson Holt <carso...@genetics.utah.edu>, Dinatale C <din...@uwindsor.ca>, "maker...@yandell-lab.org" <maker...@yandell-lab.org>
Subject: Re: [maker-devel] usage of add_utr scripts

 

Hi Anar,

 

I looks like the add_utr_to_gff3 script is completely hosed in that release.  I think it's just there as a relic and add_utr_start_stop_gff is the script that you want.  From those errors it looks like your gff3 file doesn't have the fasta sequence at the end of it. 

 

If I've guessed correctly, the quick workaround would be to just add lines like the following to the end of your GFF3 file - where ,of course, the fasta header and sequence are replaced with the correct data for the chromosome(s)/contig(s) represented in your GFF3 file. 

 

##FASTA

>chr1

ATCGATACGTAG...

 

B

 

On Sep 13, 2011, at 5:22 PM, Khan, Anar wrote:

 

From: maker-dev...@yandell-lab.org [mailto:maker-dev...@yandell-lab.org] On Behalf Of Carson Holt

To: Dinatale C; maker...@yandell-lab.org
Subject: Re: [maker-devel] usage of add_utr scripts

 

The add_utr_to_gff3 script should have been pulled out, and somehow was left in the 2.10 release.  The add_utr_start_stop_codon_gff script changes IDs, which normally isn’t too big a problem because chado will also change them (ID’s just need to be unique and are not necessarily informative).

If you need those to be unchanged, I’ll modify the script.  It will be a few days though because I’m defending my thesis on Monday.

You can also use MAKER 2.11-beta.  Named UTR is supposed to be optional in GFF3 format because it is always implicit anyway, but MAKER 2.11 now adds explicit UTR by default to better work with JBrowse.  Just give it theother MAKER’s GFF3 file and turn all the pass options to 1.  Don’t use other input files including repeatmasker options.

 

Attention: The information contained in this message and/or attachments from AgResearch Limited is intended only for the persons or entities to which it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of, or taking of any action in reliance upon, this information by persons or entities other than the intended recipients is prohibited by AgResearch Limited. If you have received this message in error, please notify the sender immediately.

 

---

 

 

<ATT00001..txt>

[Khan, Anar]

Hi

 

I’d like to try to add UTR coordinates predicted by the ab initio gene finders, given I can’t derive them from est alignments due to a lack of EST data. I see raw output is available in “the void” folder. Before writing a rather convoluted script which attempts to match snap/augustus/fgenesh predictions with maker predictions, does a script already exist; or can anyone offer suggestions/advice?

 

Cheers

Anar

 

From: maker-dev...@yandell-lab.org [mailto:maker-dev...@yandell-lab.org] On Behalf Of Khan, Anar

Sent: Thursday, 29 September 2011 2:28 p.m.
To: Carson Holt; Barry Moore

[Carson Holt]

Well… It is more complicated than you may think.  SNAP doesn't predict UTR, fgenesh couldn't predict these last time I used it (but that was a few versions ago), and Augustus can but really only for a couple of model organisms (to get it to work on other organisms you have to have the coordinates  for a significant fraction of known UTRs throughout the genome).  UTR annotation is done downstream of the ab initio prediction in most cases.  So parsing the ab initio predictions will not give you these because they really aren't part of the output to begin with.  This is why MAKER uses ESTs to predict them.

--Carson

From: "Khan, Anar" <Anar...@agresearch.co.nz>
Date: Thu, 27 Oct 2011 15:53:53 -0600
To: Carson Holt <carso...@genetics.utah.edu>, Barry Moore <bmo...@genetics.utah.edu>
Cc: "maker...@yandell-lab.org" <maker...@yandell-lab.org>, Dinatale C <din...@uwindsor.ca>
Subject: RE: [maker-devel] usage of add_utr scripts

Hi

 

I’d like to try to add UTR coordinates predicted by the ab initio gene finders, given I can’t derive them from est alignments due to a lack of EST data. I see raw output is available in “the void”folder. Before writing a rather convoluted script which attempts to match snap/augustus/fgenesh predictions with maker predictions, does a script already exist; or can anyone offer suggestions/advice?

 

Cheers

Anar

 

From: maker-dev...@yandell-lab.org [mailto:maker-dev...@yandell-lab.org] On Behalf Of Khan, Anar
Sent: Thursday, 29 September 2011 2:28 p.m.
To: Carson Holt; Barry Moore
Cc: maker...@yandell-lab.org; Dinatale C
Subject: Re: [maker-devel] usage of add_utr scripts

 

Hi Carson

 

Thanks a lot for the explanation, all clear now.

 

Cheers

Anar

 

From: Carson Holt [mailto:carso...@genetics.utah.edu]
Sent: Thursday, 29 September 2011 2:13 p.m.
To: Khan, Anar; Barry Moore
Cc: Dinatale C; maker...@yandell-lab.org
Subject: Re: [maker-devel] usage of add_utr scripts

 

ESTs that overlap the splice sites of the prediction are used to extend the gene model.  You can think of it as a kind of cut and pasteoperation.  Alt_est does not match the splice sites correctly (there are several technical reasons why).  Also since  alt_est is aligned in protein space (tblastx) the UTR region may not line up as it is non-coding, so it is not under the same evolutionary constraints as the rest of the alignment.