cuffdiff output

dnusol · 02-08-2011, 06:18 AM

Hi, while running cuffdiff on two samples (so no replicates) I get the following message

> Processing Locus Chr1:258631-267891 [ ] 0%Error: sqrt(det(cov)) == 0, 0.000000 after rounding.
> Processing Locus Chr3:2891085-2893741 [*********** ] 45%Error: sqrt(det(cov)) == 0, 0.000000 after rounding.
> Processing Locus Chr4:16101600-16102156 [****************** ] 74%Error: sqrt(det(cov)) == 0, 0.000000 after rounding.
> Processing Locus Chr5:357740-358798 [******************* ] 77%Error: sqrt(det(cov)) == 0, 0.000000 after rounding.
Error: sqrt(det(cov)) == 0, 0.000000 after rounding.

What does it mean? is it something I should worry about?

Thanks

Dave

pasta · 02-08-2011, 07:56 AM

sqrt() = calculate square root
cov() = calculate covariance
det() = ?

Basically the software cannot calculate the square root of covariance which COULD BE normal since you do not use replicates. Maybe is it doing some least square calculation. I dont think I would mind If I were you

dnusol · 02-08-2011, 11:31 PM

Hi Pasta,

thanks for your answer. I will make some fake replicates in order to check that this issue is related to the lack of replicates.

Meanwhile I run cufflinks and cuffcompare using a downloaded GTF file as reference file but the cuffdiff output does not come out with information the gene name or locus column, so I just get the random identifier generated by cufflinks. Do you know how I can include this information?? See the following example

$ head genes.fpkm_tracking
tracking_id class_code nearest_ref_id gene_short_name tss_id locus s5_FPKM s5_conf_lo s5_conf_hi s1_FPKM s1_conf_lo s1_conf_hi
XLOC_000001 - - - TSS1,TSS2 Chr1:3617-5903 5.34002 1.73637 8.94367 82.2908 64.1985 100.383
XLOC_000002 - - - ,TSS3 Chr1:6516-8763 1.09354 0.478872 1.70821 3.2086 2.41061 4.00659
XLOC_000003 - - - ,TSS4,TSS5 Chr1:23094-33226 32.8189 32.8189 32.8189 78.4804 78.4804 78.4804
XLOC_000004 - - - ,TSS6,TSS7 Chr1:50024-51202 170.472 149.429 191.516 251.281 223.25 279.312
XLOC_000005 - - - TSS8,TSS9 Chr1:52035-54631 5.50825 0.724292 10.2922 4.23105 0.110532 8.35157
XLOC_000006 - - - TSS10 Chr1:72335-74721 24.9435 18.1258 31.7612 17.9945 10.545 25.444
XLOC_000007 - - - - Chr1:82841-85073 0 0 0 5.19902 3.8691 6.52894
XLOC_000008 - - - TSS11,TSS12 Chr1:91391-95665 9.64846 3.4351 15.8618 7.29045 1.88894 12.692
XLOC_000009 - - - TSS13,TSS14,TSS15 Chr1:95851-99241 54.1341 39.4121 68.8561 57.8228 42.5951 73.0504

Best,

Dave

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Cuffdiff output - how reliable?	whoopster101	RNA Sequencing	12	02-23-2014 09:15 AM
CuffDiff output	Rachelly	Bioinformatics	11	04-17-2012 09:04 PM
Cuffdiff output	sheenams	RNA Sequencing	0	11-27-2011 04:41 PM
Cuffdiff Output Explanation	fongchun	Bioinformatics	4	11-08-2011 11:26 AM
my understanding for cuffdiff output	Huijuan	Bioinformatics	1	05-01-2011 05:42 AM

02-08-2011, 06:18 AM	#1
dnusol Senior Member Location: Spain Join Date: Jul 2009 Posts: 133	cuffdiff output Hi, while running cuffdiff on two samples (so no replicates) I get the following message > Processing Locus Chr1:258631-267891 [ ] 0%Error: sqrt(det(cov)) == 0, 0.000000 after rounding. > Processing Locus Chr3:2891085-2893741 [********* ] 45%Error: sqrt(det(cov)) == 0, 0.000000 after rounding. > Processing Locus Chr4:16101600-16102156 [************** ] 74%Error: sqrt(det(cov)) == 0, 0.000000 after rounding. > Processing Locus Chr5:357740-358798 [***************** ] 77%Error: sqrt(det(cov)) == 0, 0.000000 after rounding. Error: sqrt(det(cov)) == 0, 0.000000 after rounding. What does it mean? is it something I should worry about? Thanks Dave

02-08-2011, 07:56 AM	#2
pasta Member Location: Alps Join Date: Jan 2011 Posts: 27	sqrt() = calculate square root cov() = calculate covariance det() = ? Basically the software cannot calculate the square root of covariance which COULD BE normal since you do not use replicates. Maybe is it doing some least square calculation. I dont think I would mind If I were you

02-08-2011, 11:31 PM	#3
dnusol Senior Member Location: Spain Join Date: Jul 2009 Posts: 133	Hi Pasta, thanks for your answer. I will make some fake replicates in order to check that this issue is related to the lack of replicates. Meanwhile I run cufflinks and cuffcompare using a downloaded GTF file as reference file but the cuffdiff output does not come out with information the gene name or locus column, so I just get the random identifier generated by cufflinks. Do you know how I can include this information?? See the following example $ head genes.fpkm_tracking tracking_id class_code nearest_ref_id gene_short_name tss_id locus s5_FPKM s5_conf_lo s5_conf_hi s1_FPKM s1_conf_lo s1_conf_hi XLOC_000001 - - - TSS1,TSS2 Chr1:3617-5903 5.34002 1.73637 8.94367 82.2908 64.1985 100.383 XLOC_000002 - - - ,TSS3 Chr1:6516-8763 1.09354 0.478872 1.70821 3.2086 2.41061 4.00659 XLOC_000003 - - - ,TSS4,TSS5 Chr1:23094-33226 32.8189 32.8189 32.8189 78.4804 78.4804 78.4804 XLOC_000004 - - - ,TSS6,TSS7 Chr1:50024-51202 170.472 149.429 191.516 251.281 223.25 279.312 XLOC_000005 - - - TSS8,TSS9 Chr1:52035-54631 5.50825 0.724292 10.2922 4.23105 0.110532 8.35157 XLOC_000006 - - - TSS10 Chr1:72335-74721 24.9435 18.1258 31.7612 17.9945 10.545 25.444 XLOC_000007 - - - - Chr1:82841-85073 0 0 0 5.19902 3.8691 6.52894 XLOC_000008 - - - TSS11,TSS12 Chr1:91391-95665 9.64846 3.4351 15.8618 7.29045 1.88894 12.692 XLOC_000009 - - - TSS13,TSS14,TSS15 Chr1:95851-99241 54.1341 39.4121 68.8561 57.8228 42.5951 73.0504 Best, Dave