[R] keep track of selected observations over time

Petr Pikal petr.pikal at precheza.cz
Thu Jan 25 15:24:27 CET 2007


Hi


On 24 Jan 2007 at 18:26, Jenny persson wrote:

Date sent:      	Wed, 24 Jan 2007 18:26:17 +0100 (CET)
From:           	Jenny persson <jenny197806 at yahoo.se>
To:             	r-help at stat.math.ethz.ch
Subject:        	Re: [R] keep track of selected observations over time

> 
> Thanks Peter, I forgot that the mailinglist only accept the pdf and
> ps. file.
> 

Well, not sure if I understand what you want (your example is not 
reproducible), but a boxplot beside of making actual plot also 
returns invisibly a structure for plotting, actually a list. So you 
can call

b.str<- boxplot(.....)

and if you go through b.str you can find

$out and $group values, which indicate outliers. Then you can do e.g.

sel <- your.data[,2] %in% b.str$out[ddd$group == 1]

to get logical vector which points are marked as outliers in your 
first column.Then

your.data[sel,]

gives you appropriate rows which you can use e.g. for adding lines

lines(1:4, your.data[sel,][1,-1])

HTH
Petr

>   Here is my problem again:
> 
>   Attached is a graph  of four boxplots from one patient s data at
>   four time points, i.e. each boxplot presents the data at each time
>   point. At day 0 there are 5 extreme values from five peptide
>   sequences (please see the below data). Since the response is
>   changing over time, some of these five extreme values at day 0 may
>   be lower or higher at day 56, 112 and 252. How can I trace the
>   location of each peptide sequence that has extreme value at day 0 on
>   the box plots at day 56, 112 and 252 by color or number coding. For
>   example, the most five responding peptides can be ranked from 5
>   (highest value) to 1 (lowest), so if I do the graph again I would
>   see the five extreme values at day 0 as numbers 5-1 and each of
>   these numbers can be any where on the box plot at day 56, 112 and
>   252 or instead of the rank numbers using the peptide sequence that
>   corresponds to each value. Alternatively, the locations of each
>   peptide sequence at the four time points could be linked by a line.
>   I would like to repeat this procedure for time point 56, 112 and 252
>   as well. That is, at day 56, I want trace where on the box plot at
>   day 0, 112 and 252, each of the four peptide sequences that have
>   highest responses is. Again, these four values/peptides can be
>   presented by different colors, numbers or their peptide sequences
>   that distinguish them from the other most responding peptides at day
>   0, 112 and 252. Can I do the four procedures at the same time, I
>   mean, if at each time point I want to keep track of where the most
>   peptide responses from this time point are, then the total number of
>   peptides at the four time points could be 20. That is for each box
>   plot, there will be 20 id numbers corresponding to each peptide at
>   respectively time point.  The graph can be kind of messy.
> 
>    I have a simple solution of how to see the most responding peptides
>    changing over time, by plotting each peptide s responses at the
>    four time points. But I haven t managed the procedure above. If you
>    have any suggestion how I can do this in R, I would be very
>    thankful.
> 
>   Many thanks 
>   Jenny
> 
> 
> 
>   Part of the data:
> 
>   > pat1[1:20,]
>             peptides       P1_D0      P1_D56      P1_D112     P1_D252
>   1    AAAKKGSEQTLKS -0.06181601 -0.12610877 -0.057898384 -0.02126862
>   2    AAAAPASEDEDDE -0.10972387 -0.17174722 -0.136468783 -0.16262501
>   3    AAAAVSSAKRSLR  0.64156129  1.02630879  0.079891841  0.29757984
>   4    AAAKKGSEQESVK -0.54943062 -0.34311337 -0.338910367 -0.14526498
>   5    AAANLTKIVAYLG -1.72207627 -1.63326368 -0.459839317 -0.63302448
>   6    AACGRISYNDMFE  0.52513671  0.65123495  1.151866644  1.49481479
>   7    AAEAEKAASESLR -0.69366543 -0.47038765 -0.144156174 -0.16042556
>   8    AAEHAQSCRSSAA -0.13373130 -0.09229543 -0.102485597 -0.09782440
>   9  AAERHARLNDSYRLQ -0.19316423 -0.33164239 -0.033764989 -0.11734969
>   10   AAETISAARALPS -0.49632307 -0.53666696 -0.263024663 -0.18231712
>   11 AAEVQRFNRDVDETI -0.80014439 -0.91002202 -0.257201702 -0.12391146
>   12   AAEWTANVTAEFK -0.41544438 -0.10980658 -0.288133150 -0.32022460
>   13   AAGIQWSEEETED -0.04015673  0.08529726  0.002471231  0.07599156
>   14   AAGPALSPVPPVV -0.26795462 -0.36739148 -0.512049278 -0.25449224
>   15   AAGPPPSEGEEST -1.59272674 -1.69729759 -0.843351943 -0.49271773
>   16   AAKIASRQPDSHI -0.40722382 -0.27236225 -0.224539441 -0.32998813
>   17   AAKIQASFRGHMA  2.41234976  2.84435484  0.160852331  0.80197802
>   18   AALDLGGSSDPYV -1.21202038 -1.25109705 -0.259515922 -0.24351352
>   19   AALEPGPSESLTA -2.00256570 -1.57566020 -0.390584034 -0.23682626
>   20 AALLELWELRRQQYE  1.42797600  1.33539104  1.486154861  1.67471189
> 
> 
>   par(las=1)   # all axis labels horizontal
>         boxplot(data.frame(pat1[,c(2:5)]), pars = list(boxwex = 0.4,  
>                                      staplewex = 0.8, outwex = .5), 
>                    boxfill="lightblue",border=c(3:6), names=c("Day 0",
>                    "56 days","112 days", "252 days"),
>                                    col.main="blue",  
>                    main ="Averaged peptide response at 4 different
>                    time points for patient 200001",cex.main=0.9,
>                    font.main=0.9)
> 
> 
> 
> 
> 
> Peter Konings <peter.l.e.konings at gmail.com> skrev:
>   Dear Jenny,
> 
> Your post did not have an attachment. The mailing list software strips
> most attachments away: see the 'technical details of posting' section
> of the posting guide at: http://www.r-project.org/posting-guide.html.
> 
> HTH 
> Peter.
> 
>   On 1/24/07, Jenny persson < jenny197806 at yahoo.se > wrote:   Dear
>   all,
> 
>   Attached is a description of my data, graph and the problem which I
>   need help with. Hope you have time to open the file and help me out.
> 
>   Many thanks,
>   Jenny
> 
> 
> --------------------------------- 
> 
> 
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help 
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html and provide commented,
> minimal, self-contained, reproducible code.
> 
> 
> 
> 
> 
> 
> 
> ---------------------------------
> 
> 

Petr Pikal
petr.pikal at precheza.cz



More information about the R-help mailing list