Pindel: improved version for indels and structural variants - Page 3

ralonso · 02-29-2012, 06:00 AM

I don't have it sorted, could I use it even this fact? or is it totally necessary to sort it?

KaiYe · 02-29-2012, 07:09 AM

Quote:

Originally Posted by ralonso

I don't have it sorted, could I use it even this fact? or is it totally necessary to sort it?

It is better sorted and then let Pindel directly read your sorted bam, -i config file.

Otherwise, you may use sam2pindel to extract reads (then use -p extracted_reads), but it is not recommended, as this takes additional space and runtime.

Kai

sam.a · 03-02-2012, 01:47 AM

hi, KaiYe
i'm wondering - does Pindel works only with human genome?

KaiYe · 03-02-2012, 02:53 AM

Quote:

Originally Posted by sam.a

hi, KaiYe
i'm wondering - does Pindel works only with human genome?

Of course, as long as there is a reference genome or contigs.

mikhmv · 03-02-2012, 07:15 AM

Hi KaiYe,

I used MarkDuplicates from Picard and mark duplicates but did not removed them.
I have a question does pindel need file with removed duplicates or can work with Marked ones?

Thank you in an advance.

KaiYe · 03-02-2012, 08:01 AM

Quote:

Originally Posted by mikhmv

Hi KaiYe,

I used MarkDuplicates from Picard and mark duplicates but did not removed them.
I have a question does pindel need file with removed duplicates or can work with Marked ones?

Thank you in an advance.

It is better to remove duplicates. Pindel has an internal counting procedure to output total support and unique support.

In the original Pindel output file (not converted to vcf) per sample you will see something like:
NA18507 + 10 8 - 12 9
While 10 is total count from + strand, 8 is the unique count.
12 and 9 are for other strand, total or unique.

Kai

mikhmv · 03-02-2012, 08:48 AM

It was a quick reply!!! Thanks.

I have a several more questions now.
1. Does it mean that Pindel does not check flag that this read is duplicate?
2. If you have count it independently maybe it just doesn't matter was duplicates removed or not as a amount unique reads will be same?
3. Did pindel take in account sequencing error rate while determining amount of unique reads?

libiyagirl · 03-06-2012, 12:05 PM

Hi Kai,

I found that the bam_configuration method takes different number of reads for analysis, compared with the bam2pindel.pl method. Could you please explain it to me? And the new version, pindel024n processes way more reads compared with Pindel_0.2.4d. Does the two versions filter the reads differently?

Thank you so much for your help,

KaiYe · 03-06-2012, 01:36 PM

Quote:

Originally Posted by mikhmv

It was a quick reply!!! Thanks.

I have a several more questions now.
1. Does it mean that Pindel does not check flag that this read is duplicate?
2. If you have count it independently maybe it just doesn't matter was duplicates removed or not as a amount unique reads will be same?
3. Did pindel take in account sequencing error rate while determining amount of unique reads?

Pindel doesn't check flag for duplicate but count by itself. It reports both total and unique counts for both strands. It is up to the users to decide which number to use. Pindel considers 3% sequence error rate by default but you can change this parameter.

Kai

KaiYe · 03-06-2012, 01:39 PM

Quote:

Originally Posted by libiyagirl

Hi Kai,

I found that the bam_configuration method takes different number of reads for analysis, compared with the bam2pindel.pl method. Could you please explain it to me? And the new version, pindel024n processes way more reads compared with Pindel_0.2.4d. Does the two versions filter the reads differently?

Thank you so much for your help,

Yes, the perl program does slightly different things than Pindel internal BAM read function, although we tried to make them consistent. We changed a few parameters in pindel024n, resulting at least 3 times more candidate reads. We are investigating this impact. So far no significant difference (except for more sensitive) was observed.

Kai

mikhmv · 03-10-2012, 12:32 PM

Hi KaiYe,

I have an error message during run pindel2vcf.
The executing command was:
pindel2vcf -r $REFGENOME -R hg19b37 -d 20120310 -p 5173N_TD.res -v 5173N_TD.res2.vcf
The last few lines from program:
Processing chromosome chr11
Reading region 0-300000000
Reading chromosome chr11 into memory.
Total reads: 180
Sorting completed
Removing chromosome chr11 from memory.
Processing chromosome chr11_gl000202_random
Reading region 0-300000000
Total reads: 0
Sorting completed
Segmentation fault

Can you help to resolve this problem?

Regards

mikhmv · 03-10-2012, 01:03 PM

KaiYe,
I think it is a bug. I got this message on different chromosome depend on file but error everytime after Total reads = 0.

Is it possible to fix?

Regards,
Max

KaiYe · 03-12-2012, 03:32 AM

Quote:

Originally Posted by mikhmv

KaiYe,
I think it is a bug. I got this message on different chromosome depend on file but error everytime after Total reads = 0.

Is it possible to fix?

Regards,
Max

Hi Max,

Yes, we also noticed this bug. Please try our latest source code 0.2.4o, updated on March 9 2012.

Please do contact us if you experience the same or other issues.

Thanks,

Kai

mikhmv · 03-12-2012, 05:34 PM

Quote:

Originally Posted by KaiYe

Hi Max,

Yes, we also noticed this bug. Please try our latest source code 0.2.4o, updated on March 9 2012.

Please do contact us if you experience the same or other issues.

Thanks,

Kai

Thanks KaiYe.
"o" is working well.

mikhmv · 03-13-2012, 01:22 PM

Hi KaiYe,
I have a question how can we annotate BreakPoints? pindel2vcf doesn't support it.

Regards

mikhmv · 03-15-2012, 07:05 PM

Hi KaiYe,
Does pindel2vcf provide quality score?
How can I get it in vcf?
Regards,

bwubb · 03-22-2012, 08:29 AM

I recently upgraded to pindel024o and may have found some sort of bug; or maybe Im doing something I shouldnt be...

When I execute pindel now all of my bins are empty and it produces no results. I do not get an error, I just get no results, as such:

Code:

Pindel version 0.2.4o, March 9 2012.
Looping over all chromosomes.
Processing chromosome: 1
Chromosome Size: 249250621
NumBoxes: 60001	BoxSize: 8975

Looking at chromosome 1 bases 0 to 10000000.
There are no reads for this bin.

Looking at chromosome 1 bases 10000000 to 20000000.
There are no reads for this bin.

Looking at chromosome 1 bases 20000000 to 30000000.
There are no reads for this bin.

Looking at chromosome 1 bases 30000000 to 40000000.
There are no reads for this bin.

Looking at chromosome 1 bases 40000000 to 50000000.
There are no reads for this bin.

Looking at chromosome 1 bases 50000000 to 60000000.
There are no reads for this bin.
...

This is suppose to be exome sequencing data. Now the odd thing is that I have previous pindel024m results for this sample, so everything should work as I did not change anything. The only change besides the version is where I am executing from.

I made a dir 'my_test' in my pindel024o directory (~/pindel024o/my_test/). Previously I was in a different path, something like ~/NGS/Pindel_restults/

The BAM files used have not changed, nor have the locations. If I execute pindel from my old directory it still works.

Im just confused by this.

KaiYe · 03-22-2012, 09:00 AM

Quote:

Originally Posted by bwubb

I recently upgraded to pindel024o and may have found some sort of bug; or maybe Im doing something I shouldnt be...

When I execute pindel now all of my bins are empty and it produces no results. I do not get an error, I just get no results, as such:

Code:

Pindel version 0.2.4o, March 9 2012.
Looping over all chromosomes.
Processing chromosome: 1
Chromosome Size: 249250621
NumBoxes: 60001	BoxSize: 8975

Looking at chromosome 1 bases 0 to 10000000.
There are no reads for this bin.

Looking at chromosome 1 bases 10000000 to 20000000.
There are no reads for this bin.

Looking at chromosome 1 bases 20000000 to 30000000.
There are no reads for this bin.

Looking at chromosome 1 bases 30000000 to 40000000.
There are no reads for this bin.

Looking at chromosome 1 bases 40000000 to 50000000.
There are no reads for this bin.

Looking at chromosome 1 bases 50000000 to 60000000.
There are no reads for this bin.
...

This is suppose to be exome sequencing data. Now the odd thing is that I have previous pindel024m results for this sample, so everything should work as I did not change anything. The only change besides the version is where I am executing from.

I made a dir 'my_test' in my pindel024o directory (~/pindel024o/my_test/). Previously I was in a different path, something like ~/NGS/Pindel_restults/

The BAM files used have not changed, nor have the locations. If I execute pindel from my old directory it still works.

Im just confused by this.

This is indeed odd. Can you share a tiny BAM (first 1k reads) with me so that I can check this out. We don't see this on our data but we would like to find out. Please contact my private email address k.ye@lumc.nl to discuss how to share the data.

Thank you.

Kai

bwubb · 03-22-2012, 01:47 PM

Hi Kai,

I appreciate you offering to help with this issue, but I believe I resolved the problem though. The problem was my config file. I was not actually using the same exact config file, but had written an identical one to my 'version m' data.

Identical... except for an embarrassing mistake in that the newer config file did not have a new line character at the end.

This seemed to result in the error/results I posted above.

EWLameijer · 03-26-2012, 05:13 AM

Quote:

Originally Posted by bwubb

Hi Kai,

I appreciate you offering to help with this issue, but I believe I resolved the problem though. The problem was my config file. I was not actually using the same exact config file, but had written an identical one to my 'version m' data.

Identical... except for an embarrassing mistake in that the newer config file did not have a new line character at the end.

This seemed to result in the error/results I posted above.

Hi Bwubb,

It's very good that you resolved the problem on your own. To save you and other people some time in the future, Kai and I have now implemented some changes in the bam-file reading parts, and submitted it as Pindel-0.2.4p.

Pindel-0.2.4p does not fundamentally change the pindel output, but will give a clear error message if an user by accident:
-uses the -i option on a bam-configuration that does not exist
-uses a bam-configuration file that is empty or truncated
-refers in the bam-configuration file to bam-files that don't exist
-refers in the bam-configuration file to bam-files that have not been indexed yet

Inspired by your report, it should now also work if the last line does not end in a newline character.

Thank you for your message and help in clearing this up,

Eric-Wubbo

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Structural Variants	jkozubek	Bioinformatics	4	08-22-2016 08:38 AM
Calling structural variants (CNVs) with single-end reads	agwe	Genomic Resequencing	5	01-18-2016 08:30 AM
Strangely high proportion of variants are INDELs	PeteH	Bioinformatics	2	05-09-2012 01:13 AM
Calling structural variants from capture data	Heisman	Bioinformatics	3	04-16-2012 08:01 AM
cufflinks 1.2.0 version got me significantly different results than the old version	slowsmile	Bioinformatics	9	02-01-2012 02:26 AM

02-29-2012, 06:00 AM	#41
ralonso Member Location: Spain Join Date: Feb 2012 Posts: 10	I don't have it sorted, could I use it even this fact? or is it totally necessary to sort it? Last edited by ralonso; 02-29-2012 at 06:08 AM.

03-22-2012, 08:29 AM	#57
bwubb Member Location: Philadelphia Join Date: Jan 2012 Posts: 58	I recently upgraded to pindel024o and may have found some sort of bug; or maybe Im doing something I shouldnt be... When I execute pindel now all of my bins are empty and it produces no results. I do not get an error, I just get no results, as such: Code: Pindel version 0.2.4o, March 9 2012. Looping over all chromosomes. Processing chromosome: 1 Chromosome Size: 249250621 NumBoxes: 60001 BoxSize: 8975 Looking at chromosome 1 bases 0 to 10000000. There are no reads for this bin. Looking at chromosome 1 bases 10000000 to 20000000. There are no reads for this bin. Looking at chromosome 1 bases 20000000 to 30000000. There are no reads for this bin. Looking at chromosome 1 bases 30000000 to 40000000. There are no reads for this bin. Looking at chromosome 1 bases 40000000 to 50000000. There are no reads for this bin. Looking at chromosome 1 bases 50000000 to 60000000. There are no reads for this bin. ... This is suppose to be exome sequencing data. Now the odd thing is that I have previous pindel024m results for this sample, so everything should work as I did not change anything. The only change besides the version is where I am executing from. I made a dir 'my_test' in my pindel024o directory (~/pindel024o/my_test/). Previously I was in a different path, something like ~/NGS/Pindel_restults/ The BAM files used have not changed, nor have the locations. If I execute pindel from my old directory it still works. Im just confused by this.

03-02-2012, 01:47 AM	#43
sam.a Junior Member Location: Poland Join Date: Nov 2011 Posts: 1	hi, KaiYe i'm wondering - does Pindel works only with human genome?

03-02-2012, 07:15 AM	#45
mikhmv Registered Vendor Location: MD Join Date: Feb 2012 Posts: 18	Hi KaiYe, I used MarkDuplicates from Picard and mark duplicates but did not removed them. I have a question does pindel need file with removed duplicates or can work with Marked ones? Thank you in an advance.

03-02-2012, 08:48 AM	#47
mikhmv Registered Vendor Location: MD Join Date: Feb 2012 Posts: 18	It was a quick reply!!! Thanks. I have a several more questions now. 1. Does it mean that Pindel does not check flag that this read is duplicate? 2. If you have count it independently maybe it just doesn't matter was duplicates removed or not as a amount unique reads will be same? 3. Did pindel take in account sequencing error rate while determining amount of unique reads?

03-06-2012, 12:05 PM	#48
libiyagirl Member Location: canada Join Date: Jan 2011 Posts: 27	Hi Kai, I found that the bam_configuration method takes different number of reads for analysis, compared with the bam2pindel.pl method. Could you please explain it to me? And the new version, pindel024n processes way more reads compared with Pindel_0.2.4d. Does the two versions filter the reads differently? Thank you so much for your help,

03-10-2012, 12:32 PM	#51
mikhmv Registered Vendor Location: MD Join Date: Feb 2012 Posts: 18	Hi KaiYe, I have an error message during run pindel2vcf. The executing command was: pindel2vcf -r $REFGENOME -R hg19b37 -d 20120310 -p 5173N_TD.res -v 5173N_TD.res2.vcf The last few lines from program: Processing chromosome chr11 Reading region 0-300000000 Reading chromosome chr11 into memory. Total reads: 180 Sorting completed Removing chromosome chr11 from memory. Processing chromosome chr11_gl000202_random Reading region 0-300000000 Total reads: 0 Sorting completed Segmentation fault Can you help to resolve this problem? Regards

03-10-2012, 01:03 PM	#52
mikhmv Registered Vendor Location: MD Join Date: Feb 2012 Posts: 18	KaiYe, I think it is a bug. I got this message on different chromosome depend on file but error everytime after Total reads = 0. Is it possible to fix? Regards, Max

03-13-2012, 01:22 PM	#55
mikhmv Registered Vendor Location: MD Join Date: Feb 2012 Posts: 18	Hi KaiYe, I have a question how can we annotate BreakPoints? pindel2vcf doesn't support it. Regards

03-15-2012, 07:05 PM	#56
mikhmv Registered Vendor Location: MD Join Date: Feb 2012 Posts: 18	Hi KaiYe, Does pindel2vcf provide quality score? How can I get it in vcf? Regards,

03-22-2012, 01:47 PM	#59
bwubb Member Location: Philadelphia Join Date: Jan 2012 Posts: 58	Hi Kai, I appreciate you offering to help with this issue, but I believe I resolved the problem though. The problem was my config file. I was not actually using the same exact config file, but had written an identical one to my 'version m' data. Identical... except for an embarrassing mistake in that the newer config file did not have a new line character at the end. This seemed to result in the error/results I posted above.