Synapter - makeMaster() and synergise() functions warnings/fails
6
0
Entering edit mode
sogueta • 0
@sogueta-7100
Last seen 9.3 years ago
Spain

Hi everyone

Please, I need some explanations about the following warning messages and fails I have obtained after testing synapter package with my own samples:

First, I have generated a master file:

> cmb <- estimateMasterFdr(hdmse, fastadb, masterFdr=0.05, verbose=TRUE)
12 Peptide files available - 4083 combinations
Processing HDMSe_PlasmaC25_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaC31_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaC32_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaI13_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaI16_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaI17_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaII03_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaII07_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaII14_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaIII01_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaIII02_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaIII04_10102014_IA_final_peptide.csv.gz
Generating unique proteotypic peptides...
Calculating...
>
> master <- makeMaster(hdmse[bestComb(cmb)], verbose=TRUE)
Processing HDMSe_PlasmaC25_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaC31_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaC32_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaI13_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaI16_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaI17_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaII03_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaII14_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaIII01_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaIII02_10102014_IA_final_peptide.csv.gz
Processing HDMSe_PlasmaIII04_10102014_IA_final_peptide.csv.gz
Master: HDMSe_PlasmaII14_10102014_IA_final_peptide.csv.gz (1883 peptides)
 |--- (1) Merged: 1883 features.
 +- Merging master and HDMSe_PlasmaIII02_10102014_IA_final_peptide.csv.gz (1878 peptides)
 |--- (2) Merged: 2320 features.
 +- Merging master and HDMSe_PlasmaIII01_10102014_IA_final_peptide.csv.gz (1829 peptides)
 |--- (3) Merged: 2632 features.
 +- Merging master and HDMSe_PlasmaIII04_10102014_IA_final_peptide.csv.gz (1453 peptides)
 |--- (4) Merged: 2729 features.
 +- Merging master and HDMSe_PlasmaII03_10102014_IA_final_peptide.csv.gz (1430 peptides)
 |--- (5) Merged: 2806 features.
 +- Merging master and HDMSe_PlasmaI13_10102014_IA_final_peptide.csv.gz (1331 peptides)
 |--- (6) Merged: 2908 features.
 +- Merging master and HDMSe_PlasmaI17_10102014_IA_final_peptide.csv.gz (1170 peptides)
 |--- (7) Merged: 2950 features.
 +- Merging master and HDMSe_PlasmaC25_10102014_IA_final_peptide.csv.gz (1136 peptides)
 |--- (8) Merged: 3019 features.
 +- Merging master and HDMSe_PlasmaI16_10102014_IA_final_peptide.csv.gz (1115 peptides)
 |--- (9) Merged: 3079 features.
 +- Merging master and HDMSe_PlasmaC32_10102014_IA_final_peptide.csv.gz (1031 peptides)
 |--- (10) Merged: 3118 features.
 +- Merging master and HDMSe_PlasmaC31_10102014_IA_final_peptide.csv.gz (432 peptides)
 \--- (11) Merged: 3144 features.
Master: HDMSe_PlasmaIII02_10102014_IA_final_peptide.csv.gz (1878 peptides)
 |--- (1) Merged: 1878 features.
 +- Merging master and HDMSe_PlasmaIII01_10102014_IA_final_peptide.csv.gz (1829 peptides)
 |--- (2) Merged: 2363 features.
 +- Merging master and HDMSe_PlasmaIII04_10102014_IA_final_peptide.csv.gz (1453 peptides)
 |--- (3) Merged: 2499 features.
 +- Merging master and HDMSe_PlasmaII03_10102014_IA_final_peptide.csv.gz (1430 peptides)
 |--- (4) Merged: 2618 features.
 +- Merging master and HDMSe_PlasmaI13_10102014_IA_final_peptide.csv.gz (1331 peptides)
 |--- (5) Merged: 2741 features.
 +- Merging master and HDMSe_PlasmaI17_10102014_IA_final_peptide.csv.gz (1170 peptides)
 |--- (6) Merged: 2790 features.
 +- Merging master and HDMSe_PlasmaC25_10102014_IA_final_peptide.csv.gz (1136 peptides)
 |--- (7) Merged: 2871 features.
 +- Merging master and HDMSe_PlasmaI16_10102014_IA_final_peptide.csv.gz (1115 peptides)
 |--- (8) Merged: 2936 features.
 +- Merging master and HDMSe_PlasmaC32_10102014_IA_final_peptide.csv.gz (1031 peptides)
 |--- (9) Merged: 2980 features.
 +- Merging master and HDMSe_PlasmaC31_10102014_IA_final_peptide.csv.gz (432 peptides)
 |--- (10) Merged: 3007 features.
 +- Merging master and HDMSe_PlasmaII14_10102014_IA_final_peptide.csv.gz (1883 peptides)
 \--- (11) Merged: 3148 features.
Warning messages:
1: In simpleLoess(y, x, w, span, degree, parametric, drop.square, normalize,  :
  pseudoinverse used at 48.645
2: In simpleLoess(y, x, w, span, degree, parametric, drop.square, normalize,  :
  neighborhood radius 0.1093
3: In simpleLoess(y, x, w, span, degree, parametric, drop.square, normalize,  :
  reciprocal condition number  5.482e-017
4: In simpleLoess(y, x, w, span, degree, parametric, drop.square, normalize,  :
  There are other near singularities as well. 0.011946

Which is the meaning of all those warning messages?

After master file was created, I would like to test the synergise() function with one HDMSe/MSe sample typing the following:

> prueba <- synergise(filenames = list(
+                       identpeptide = hdmsemaster,
+                       quantpeptide = mse[1],
+                       quantpep3d = mse3d[1],
+                       fasta = fastadb),
+                     master = TRUE,
+                     outputdir = tempdir(),
+                     grid.nsd.from = 1, grid.nsd.to = 6, grid.nsd.by = 1,
+                     grid.ppm.from = 7, grid.ppm.to = 11, grid.ppm.by = 1,
+                     grid.subset = 0.1, uniquepep = TRUE)

and the result I have obtained is:

Reading quantitation final peptide file...
Reading master identification peptide file...
Reading quantitation Pep3D file...
Error in xx$loadMasterData() :
  The Pep3D file ‘C:/tmp/Synapter/Datos/MSe_PlasmaC25_MARSHSA_10102014_Pep3DAMRT.csv.gz’ does not correspond to the given Quantitation Final Peptide file ‘C:/tmp/Synapter/Datos/MSe_PlasmaC25_MARSHSA_10102014_IA_final_peptide.csv.gz’!

 

Both, final_peptide file and Pep3DAMRT file were created at the same time with PLGS software so I do not understand why synergise is saying that.

Any suggestion/advice/explanation??

Thanks in advance

Best regards

Samuel

 

software error warnings synapter • 2.3k views
ADD COMMENT
1
Entering edit mode
@laurent-gatto-5645
Last seen 11 weeks ago
Belgium

Dear Samuel,

The warnings during the master creation do not point to anything nasty, I believe. Just make sure that the retention time model in the final report looks sensible. You could explore different `span` values, that are passed to the underlying loess function.

The error results in a mismatch between Pep3D file and Quantitation Final Peptide. This is detailed here: the test does not look at the file names, but at the EMRT identifiers, but I agree that it is surprising. Could you check the following:

idx <- match(quant$precursor.leID, pep3d$spectrumID)
is.na(idx)
table(quant$precursor.inten == pep3d$Counts[idx])

where quant and pep3d are the data.frames from these two files.

ADD COMMENT
1
Entering edit mode
@sebastian-gibb-6858
Last seen 5.5 years ago
Germany

As described here we recognized the problem some time ago. Because of that I assume that there is no problem with the IDs but with the corresponding intensities. It sounds suprising but in some cases the intensity values in a final_peptide file and its corresponding Pep3D file don't match. A simple (but ugly) workaround would be to change the function on the fly:

fixInNamespace("isCorrespondingPep3DataFile", "synapter")

you will get an editor and replace it with:

function (quant, pep3d) { return(TRUE) }

This will change isCorrespondingPep3DataFile for the running R session and disable the file check. We will investigate the problem.

ADD COMMENT
0
Entering edit mode

Samuel, we will provide a more elegant fix later tonight. I will also fix this in the current release.

ADD REPLY
0
Entering edit mode

OK. I am waiting for that elegant fix :-D

In the meanwhile, I am going to play with an editor....

Thanks!!!

ADD REPLY
0
Entering edit mode
sogueta • 0
@sogueta-7100
Last seen 9.3 years ago
Spain

Hi Laurent.

Check done.

Output for is.na(idx), thousands of FALSE

Output for table(quant$precursor.inten == pep3d$Counts[idx]):

FALSE  TRUE
   18      53001

Could quant$precursor.inten be quant$precursor.inten[idx] as well? If true, the result changes to:

FALSE  TRUE
 4197     1

 

Regards!!

ADD COMMENT
0
Entering edit mode

table(quant$precursor.inten == pep3d$Counts[idx]) should be all TRUE

Sorry, could you check that allis.na(idx)) is FALSE.

ADD REPLY
0
Entering edit mode

Yes, all of then are FALSE

ADD REPLY
0
Entering edit mode
sogueta • 0
@sogueta-7100
Last seen 9.3 years ago
Spain

Hi again Laurent

Regarding the makeMaster functions and its warnings, as you suggested, changing the span value from 0.05 to 0.1 all those warnings have disapeared. Thanks!!

ADD COMMENT
0
Entering edit mode

We tried to use sensible default values, that worked well for all/most of our data, but there might always be some tuning necessary. Just make sure to check the actual alignment in the report later on. There is an illustration of the parameter in ?Synapter.

ADD REPLY
0
Entering edit mode
@laurent-gatto-5645
Last seen 11 weeks ago
Belgium

This should be fixed in devel `1.9.3` and release `1.8.3` and become available with `biocLite` in about 24 hours. 

Update: This issue has been observed with PLGS 2.5.2 and has apparently been fixed in PLGS 3.0. The latest synapter should cope whatever PLGS version.

ADD COMMENT
0
Entering edit mode

Thank Laurent. As soon it becomes available, I will install those new versions and keep you updated.

Regards.

ADD REPLY
0
Entering edit mode
sogueta • 0
@sogueta-7100
Last seen 9.3 years ago
Spain

Thanks Laurent and Sebastian

I have been working with the new and corrected version of synapter and now it runs fine, almost with no fatal errors!

Now, I am going to "fight" with R and your script to fit it to my particulate analysis. Probably, I will need some help from you.

Good job. Thanks again and best regards.

ADD COMMENT

Login before adding your answer.

Traffic: 950 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6