Problem with cummeRbund and RSQLite

[noplacetoh1de]

Hi! I'm using the Cufflinks protocol published in Nature (Vol.7 No.3, 2012) and facing a problem with cummeRbund while processing my results. Something about the SQL-Statement seems to be wrong. Here are the commands, I used:

> library(cummeRbund)
> cuff<-readCufflinks()
Creating database /media/md0/rnaseq/cuffData.db
Reading /media/md0/rnaseq/genes.fpkm_tracking
Checking samples table...
Populating samples table...
Writing genes table
Reshaping geneData table
Recasting
Writing geneData table
Error in sqliteExecStatement(con, statement, bind.data) : 
  RS-DBI driver: (unable to bind data for parameter ':status')

Already tried version 1.0.0 and 1.1.5 without any success.

Although the error is generated, the database file: cuffData.db has been created. While using the read-command again I get a CuffSet-Instance, but it seems to be incomplete:

CuffSet instance with:
	 1 samples
	 116962 genes
	 0 isoforms
	 0 TSS
	 0 CDS
	 0 promoters
	 0 splicing
	 0 relCDS

I have used cuffdiff on 6 samples in 2 conditions. Could there occur any problems with the sample/condition-description, the one has a '+' and the other a '-' ?!?

Many thanks in advance!

Lukas

[lgoff]

Hi Lukas,
What version of cufflinks/cuffdiff are you using or did you use to build the results files? There were some changes to cufflinks (particularly the layout of the 'status' columns, that affect the parsing by cummeRbund. If you are using a stale version of cufflinks, can you try to re-run cuffdiff with the most recent version of cuffdiff and try to load again.

Also, if you need to rebuild the cuffData.db, you can use the 'rebuild=T' argument to readCufflinks()

cuff<-readCufflinks(rebuild=T)

If this doesn't help, please email me directly and I'll help you get it up and running.

Cheers,
Loyal

[bruce01]

Hi,

I am getting same error, cufflinks/diff are 1.3.0 which seems to be most up-to-date versions. I get the error but db is not created, then if I rerun the command to create db (readCufflinks) I don't get the error but I only have one sample which is not the case.

Any help much appreciated, commands below,

Bruce.


cuffdiff -o bam_cuffDiff -b /data/shared/genomes/UMD_3.1/Bos_taurus.UMD3.1.64.fa -p 12 -L FertP,FertN -u ./merged_asm/merged.gtf ./BAM/E_1,./BAM/E_3,./BAM/E_4,./BAM/E_5,./BAM/E_7,./BAM/E_22,./BAM/E_25 ./BAM/E_8,./BAM/E_11,./BAM/E_12,./BAM/E_13,./BAM/E_17,./BAM/E_18,./BAM/E_21


>cuff_data<-readCufflinks("bam_cuffDiff")
Creating database bam_cuffDiff/cuffData.db
Reading bam_cuffDiff/genes.fpkm_tracking
Checking samples table...
Populating samples table...
Writing genes table
Reshaping geneData table
Recasting
Writing geneData table
Error in sqliteExecStatement(con, statement, bind.data) :
RS-DBI driver: (unable to bind data for parameter ':status')

> cuff_data
Error: object 'cuff_data' not found

> cuff_data<-readCufflinks("bam_cuffDiff")
> cuff_data
CuffSet instance with:
1 samples
24493 genes
0 isoforms
0 TSS
0 CDS
0 promoters
0 splicing
0 relCDS

> sessionInfo()
R version 2.15.0 (2012-03-30)
Platform: i386-pc-mingw32/i386 (32-bit)

locale:

[1] LC_COLLATE=English_Ireland.1252 LC_CTYPE=English_Ireland.1252
[3] LC_MONETARY=English_Ireland.1252 LC_NUMERIC=C
[5] LC_TIME=English_Ireland.1252

attached base packages:

[1] stats graphics grDevices utils datasets methods base
other attached packages:

[1] cummeRbund_1.2.0 reshape2_1.2.1 ggplot2_0.9.0 RSQLite_0.11.1
[5] DBI_0.2-5

loaded via a namespace (and not attached):

[1] colorspace_1.1-1 dichromat_1.2-4 digest_0.5.2 grid_2.15.0
[5] MASS_7.3-17 memoise_0.1 munsell_0.3 plyr_1.7.1
[9] proto_0.3-9.2 RColorBrewer_1.0-5 scales_0.2.0 stringr_0.6

[alessandra85]

Hi Lukas,

Could you resolve your problem. I am getting the same error? I appreciate if you can help with that. I am using cuff/diff 1.3.0.

Thanks

[lgoff]

Hi all. In most cases, this is a result of using a sample name or names that R does not treat as a character string. For example, if you have a sample named 'T' for treatment, R will assume that this is a logical value ('T' = true) and treat it differently. R also does not allow certain characters to be colum names in files (eg. +,-,*,%,etc) and this can have some adverse effects as well. I am working on a fix for this in cummerbund, but in the meantime, my advice is to re-run cuffdiff and us descriptive, alpha-numerical (although not beginning with a number) sample names as labels.

Cheers,
Loyal

[jiangs]

Hi，
I really need your help!
My R version is 3.0.1, and when I type"cuff_data < - readCufflinks(rebuild=T)", it shows the error:
Error in sqliteExecStatement(con, statement, bind.data) : RS-DBI driver: (error in statement: disk I/O error)
Why is that?

[laurenbhorowitz]

Having same problem

Hi,
was this problem ever resolved. I'm having a very similar issue when I try to run readCufflinks. I get the average FPKM values for my samples, but cummerbund for some reason doesn't extract any information about each replicate, nor about the amount of isoforms, TSS, promoters, etc.. Do you think this could have something to do with the fact that I didn't look for new transcripts during this run and instead ran my samples from tophat --> cuffquant --> cuffdiff using a reference transcriptome?
Thanks

[Cytosine]

Anyone yet tried updating their SQLite and RSQL-DBI?

Seems like the problem is in the database interface.

[laurenbhorowitz]

So I think I may actually have a number of issues. For your information I ran cuffdiff twice - once without new transcript discovery (I ran tophat --> cuffquant --> cuffdiff), and the second time with new transcript discover (tophat --> cufflinks --> cuffmerge --> cuffquant --> cuffdiff).
My problems are:
1. whenever I try to load in cummerbund my first data set from cuffdiff without new transcript discovery I always get a cuffset that only recognizes the genes, but doesn't recognize the isoforms, promoters, etc.. as shown below:
> cuff_data <- readCufflinks('diff_out_2c',gtfFile=genes.gtf,genome=genome.fa)
> cuff_data
CuffSet instance with:
4 samples
39677 genes
0 isoforms
0 TSS
0 CDS
0 promoters
0 splicing
0 relCDS

I know that that information was extracted because I have viewed those output files from cuffdiff separately in another program.
2. Additionally, for the first data set, cummeRbund also doesn't recognize any of the replicates for my samples. For example, cummeRbund will only report the mean FPKM value for a sample but cannot recognize the FPKM values for the individual replicates. Again, I know that those values are there because I have opened them in excel and have seen the individual values.

Does anyone have any idea what can be going on here?
Does cummeRbund not work with cuffdiff samples that didn't use new transcript discover?

[lgoff]

Hi all,
I think i've patched cummeRbund for RSQLite 1.0.0 compatibility. It now builds and runs on my machine. I pushed it up to BioC, but it may take a few days to get through the nightly build system.

If you would like to try and build it yourself, you can find the patched version (and the current development version) here:

https://github.com/gofflab/cummeRbund

You will have to:
1) clone the repository
2) R CMD BUILD cummeRbund
3) R CMD INSTALL cummeRbund

Please make sure that you have all the required dependencies installed (including RSQLite 1.0.0). The BioC installer checks this for you, but if you build yourself, it will fail if you do not have all the dependencies installed.

Also, please let me know if you continue to have errors. I don't check seqanswers all that often anymore due to time constraints, but I'll try to keep up on this specific thread.

Cheers,
Loyal