XHMM and Ion Torrent
Hi,
I'm trying to use XHMM to detect CNV with BAM files from PGM Ion Torrent. I'm using the guide on http://atgu.mgh.harvard.edu/xhmm/tutorial.shtml but I have some problems.
1st
I tried to use the example files to learn the pipeline and I used the reference genome hs37 (I downloaded it from https://ftp.1000genomes.ebi.ac.uk/vo...embly_sequence) and I obteined the output with success. So I tried to use XHMM with my BAM (2 group of 6 and 7 BAM each) files and for the first step (coverage and depth analysis) I can't use the hs37 reference genome because the program return this error:
##### ERROR MESSAGE: Badly formed genome location: Contig 'chr17' does not match any contig in the GATK sequence dictionary derived from the reference; are you sure you are using the correct reference fasta file?
For this step I tried to use two different exome.interval_list file in 2 different format:
1) 13:32900238-32900287
2) chr13:32889617-32889804
but in both of cases the error was the same.
So I tried to use GRCh37.p13.genome for reference genome with the exome list in format chr13:32889617-32889804. Is it correct?
2nd
If I use the GRCh37.p13 reference genome I can continue the analysis but when I have to do the
"Filters samples and targets and then mean-centers the targets" step if I use the parameters on the guide I get an empty file, so I tried to modify the parameters, in particular "maxMeanSampleRD" but even if I see my positive patient (with the CNV) in the file (DATA.filtered_centered.RD.txt) my final output file (DATA.xcnv) is empty (so I can't see my CNV). How can I resolve? Where am I wrong?
3rd
I see that the example BAM files are from Illumina but I use PGM Ion Torrent with amplicon's kit. Is it a problem for the analysis?
I make so many attempts and, if I don't wrong, my problem is on two particular filters: target and sample. Is there any hint for this problem?
I apologize for the questions but this is my first time with the real work as a bioinformatic so I'm so inexperienced.
Thank you in advance for the answers.
I'm trying to use XHMM to detect CNV with BAM files from PGM Ion Torrent. I'm using the guide on http://atgu.mgh.harvard.edu/xhmm/tutorial.shtml but I have some problems.
1st
I tried to use the example files to learn the pipeline and I used the reference genome hs37 (I downloaded it from https://ftp.1000genomes.ebi.ac.uk/vo...embly_sequence) and I obteined the output with success. So I tried to use XHMM with my BAM (2 group of 6 and 7 BAM each) files and for the first step (coverage and depth analysis) I can't use the hs37 reference genome because the program return this error:
##### ERROR MESSAGE: Badly formed genome location: Contig 'chr17' does not match any contig in the GATK sequence dictionary derived from the reference; are you sure you are using the correct reference fasta file?
For this step I tried to use two different exome.interval_list file in 2 different format:
1) 13:32900238-32900287
2) chr13:32889617-32889804
but in both of cases the error was the same.
So I tried to use GRCh37.p13.genome for reference genome with the exome list in format chr13:32889617-32889804. Is it correct?
2nd
If I use the GRCh37.p13 reference genome I can continue the analysis but when I have to do the
"Filters samples and targets and then mean-centers the targets" step if I use the parameters on the guide I get an empty file, so I tried to modify the parameters, in particular "maxMeanSampleRD" but even if I see my positive patient (with the CNV) in the file (DATA.filtered_centered.RD.txt) my final output file (DATA.xcnv) is empty (so I can't see my CNV). How can I resolve? Where am I wrong?
3rd
I see that the example BAM files are from Illumina but I use PGM Ion Torrent with amplicon's kit. Is it a problem for the analysis?
I make so many attempts and, if I don't wrong, my problem is on two particular filters: target and sample. Is there any hint for this problem?
I apologize for the questions but this is my first time with the real work as a bioinformatic so I'm so inexperienced.
Thank you in advance for the answers.
Comment