Results form R are different respect online website

[michela12345]

When I perform xCell online throug the website i have results like this:

<style> </style>

	X1_11S_A	X1_20_A	X1_20_C	X1_22_A	X1_22_C	X1_25_A
Adipocytes	0.0184	0.0464	0.0481	0.0357	0.0346	0.0444
Astrocytes	0.0052	0	0	0	0.0099	0
B-cells	0.0356	0	0	0.0056	0.0139	0.0192
Basophils	0.049	0.1879	0.201	0.1506	0.0749	0.1558

`

but when i do it on R i have this:

Adipocytes | 0.000000e+00 | 0.000000e+00 | 1.135611e-19 | 0.000000e+00

Astrocytes | 0.000000e+00 | 3.940393e-18 | 0.000000e+00

B-cells | 0.000000e+00 | 0.000000e+00

Basophils | 1.655104e-02 | 7.256572e-02 | 9.093160e-02 | 5.613090e-02

Do you think are the same? How I can convert the number on R?

And, when I run it online, i can download a p.value file but i do not understand what it is, can you explain me?

Thank you
Michela

[dviraran]

The online version uses rounding. You can try it your self -

scores = round(scores*1000)/1000

P-values are an attempt to assess whether the scores are significantly different than zero. If the p-values are high (>0.2) it means that the presence of this cell type in the mixture is low. I should note that providing confidence here is tough, so I won't make much of it.

Best,
Dvir

[michela12345]

I round it as you say but the results are still different. My passages are:

devtools::install_github('dviraran/xCell',force = TRUE)
library(xCell)
xcell_r <- xCellAnalysis(d0)
xcell_r = round(xcell_r*1000)/1000

am i doing something wrong? do you think thant is better the version online or on R?

[michela12345]

I used the same matrix for both the analysis, and on the online website i condiser N=64 and I click on the RNA-seq bottom because my matrix is a bulk RNA-seq. My matrix looks like this

( the data are on excel so they have a , but when I upload the file on the website I use a txt file)

[dviraran]

Hi Michela,

I can't guess what is the issue, but you sent me the data and I can see from it that there are differences in the names of the samples and the order, so my intuition is that you are not using the same TPM matrix for both analyses. Beyond that I can't tell you what is going on. I tried myself one dataset and there where no differences.

I strongly recommend using the R version, with limited cell types which are relevant for your cell types.

https://link.springer.com/protocol/10.1007/978-1-0716-0327-7_19

Best,
Dvir

[michela12345]

Ok thank you! Best, Michela Il giorno mar 9 gen 2024 alle ore 13:15 dviraran ***@***.***> ha scritto:

…

Hi Michela, I can't guess what is the issue, but you sent me the data and I can see from it that there are differences in the names of the samples and the order, so my intuition is that you are not using the same TPM matrix for both analyses. Beyond that I can't tell you what is going on. I tried myself one dataset and there where no differences. I strongly recommend using the R version, with limited cell types which are relevant for your cell types. https://link.springer.com/protocol/10.1007/978-1-0716-0327-7_19 Best, Dvir — Reply to this email directly, view it on GitHub <#81 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AU42ONHOWYMELCU6AHXPX63YNUYGVAVCNFSM6AAAAABBRGHT7KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOBSHE3DKOJTHA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[michela12345]

There is a way to say to R that my matrix is a microarray and not a bulk RNA-seq? Il giorno mar 9 gen 2024 alle ore 14:04 Michela Francesconi < ***@***.***> ha scritto:

…

Ok thank you! Best, Michela Il giorno mar 9 gen 2024 alle ore 13:15 dviraran ***@***.***> ha scritto: > Hi Michela, > > I can't guess what is the issue, but you sent me the data and I can see > from it that there are differences in the names of the samples and the > order, so my intuition is that you are not using the same TPM matrix for > both analyses. Beyond that I can't tell you what is going on. I tried > myself one dataset and there where no differences. > > I strongly recommend using the R version, with limited cell types which > are relevant for your cell types. > > https://link.springer.com/protocol/10.1007/978-1-0716-0327-7_19 > > Best, > Dvir > > — > Reply to this email directly, view it on GitHub > <#81 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AU42ONHOWYMELCU6AHXPX63YNUYGVAVCNFSM6AAAAABBRGHT7KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQOBSHE3DKOJTHA> > . > You are receiving this because you authored the thread.Message ID: > ***@***.***> >