Skip to main content ↑ ↓
Topic: different versions R influence the result of XCMS grouping?? (Read 5299 times) previous topic - next topic

different versions R influence the result of XCMS grouping??

Hi everyone,
I meet a very very strange situation!
I use XCMS for peak detection in R version 3.2.2. Then I try to use the IPO package to optimize the grouping parameters. However, IPO is not available for R 3.2.2. In this case, I change my R to 3.0.3 version. Here, I get the optimized parameters of bw and mzwid after IPO.
By using exactly same xcms object, as shown below:
> hilicNEG_XS_re                              
An "xcmsSet" object with 152 samples                              
Time range: 3-749.6 seconds (0.1-12.5 minutes)                              
Mass range: 34.9684-1599.6728 m/z                              
Peaks: 1064740 (about 7005 per sample)                              
Peak Groups: 0                               
Sample classes: Batch_A.QA, Batch_B.QB, Batch_C.QC, Batch_D.QD, Batch_E.QE, Batch_F.QF, Batch_G.QG, Batch_H.QH, Batch_A.Ref, Batch_B.Ref, Batch_C.Ref, Batch_D.Ref, Batch_E.Ref, Batch_F.Ref, Batch_G.Ref, Batch_H.Ref, Batch_A.SA, Batch_B.SB, Batch_C.SC, Batch_D.SD, Batch_E.SE, Batch_F.SF, Batch_G.SG, Batch_H.SH                               
                              
Profile settings: method = bin                              
                  step = 0.1                              
                              
Memory usage: 92.2 MB
If I DO GROUPING IN R VERSION 3.0.3, I have only 4850 groups, but I get Peak Groups: 16490 when I use VERSION 3.2.2!!!!!!!!!!! :shock:  :shock:  :shock:  :shock:  :shock:  :shock:
1. VERSION 3.0.3
An "xcmsSet" object with 152 samples                              
                              
Time range: 3-749.6 seconds (0.1-12.5 minutes)                              
Mass range: 34.9684-1599.6728 m/z                              
Peaks: 1064740 (about 7005 per sample)                              
Peak Groups: 4850                              
Sample classes: Batch_A.QA, Batch_B.QB, Batch_C.QC, Batch_D.QD, Batch_E.QE, Batch_F.QF, Batch_G.QG, Batch_H.QH, Batch_A.Ref, Batch_B.Ref, Batch_C.Ref, Batch_D.Ref, Batch_E.Ref, Batch_F.Ref, Batch_G.Ref, Batch_H.Ref, Batch_A.SA, Batch_B.SB, Batch_C.SC, Batch_D.SD, Batch_E.SE, Batch_F.SF, Batch_G.SG, Batch_H.SH                               
                              
Profile settings: method = bin                              
                  step = 0.1                              
                              
Memory usage: 100 MB   
2. R VERSION 3.2.2
An "xcmsSet" object with 152 samples                              
Time range: 3-749.6 seconds (0.1-12.5 minutes)                              
Mass range: 34.9684-1599.6728 m/z                              
Peaks: 1064740 (about 7005 per sample)                              
Peak Groups: 16490                               
Sample classes: Batch_A.QA, Batch_B.QB, Batch_C.QC, Batch_D.QD, Batch_E.QE, Batch_F.QF, Batch_G.QG, Batch_H.QH, Batch_A.Ref, Batch_B.Ref, Batch_C.Ref, Batch_D.Ref, Batch_E.Ref, Batch_F.Ref, Batch_G.Ref, Batch_H.Ref, Batch_A.SA, Batch_B.SB, Batch_C.SC, Batch_D.SD, Batch_E.SE, Batch_F.SF, Batch_G.SG, Batch_H.SH                                                       
Profile settings: method = bin                              
                  step = 0.1                                                            
Memory usage: 100 MB               

What is going on here??

Looking forward your reply!!!!