Procrustes analysis

[Prasanth S]

Hi,

We have two distance matrices for 6 subjects collected from "n" and "0"  (say) .Our requirement is to create a 3D plots for procrustes analysis from this matrices using emperor.
we only have these matrices data available. We don't have any data for Barcodesequence, LinkerPrimerSequence,DOB,Treatment & Description. How we can proceed with this data to perform procrustes analysis.? Please help

Matrix_1 for "n":

    1_n    2_n    3_n    4_n    5_n    6_n
1_n    0    0.3571579109    0.2243945249    0.3497900415    0.4904549333    0.339626282
2_n    0.3571579109    0    0.3606127292    0.4474355808    0.5343987341    0.3855148557
3_n    0.2243945249    0.3606127292    0    0.2543794679    0.3797520547    0.2933860357
4_n    0.3497900415    0.4474355808    0.2543794679    0    0.2300678446    0.2745535548
5_n    0.4904549333    0.5343987341    0.3797520547    0.2300678446    0    0.2930452708
6_n    0.339626282    0.3855148557    0.2933860357    0.2745535548    0.2930452708    0


Matrix_2  for "O":

    1_0    2_0    3_0    4_0    5_0    6_0
1_0    0    0.0272716574    0.0296064573    0.5642054462    0.4580837086    0.5611546895
2_0    0.0272716574    0    0.0168651113    0.5755955941    0.4654149449    0.5660451122
3_0    0.0296064573    0.0168651113    0    0.5707971681    0.4602898962    0.5627154804
4_0    0.5642054462    0.5755955941    0.5707971681    0    0.2439560879    0.2858049466
5_0    0.4580837086    0.4654149449    0.4602898962    0.2439560879    0    0.2693592553
6_0    0.5611546895    0.5660451122    0.5627154804    0.2858049466    0.2693592553    0

[Jenya Kopylov]

Hi Prasanth,

You may find this tutorial useful for Procrustes analysis in QIIME here?

Otherwise, as mentioned in this post, you can leave the columns for the mentioned metadata fields empty (though the column name should exist) and visualization should work.

Let us know if Emperor throws any errors,

Jenya

[Prasanth S]

Hi Jenya Kopylov,

Thank you for your response.
we already did like this but we can't able to identify the two different samples from the plot.
SInce the samples in the two matrices are different the emperor is showing error.We created a sample_ID map as follows to compare the subjects from two different matrices.

Sample_id

1_n =1
2_n=2
...

1_0=1
2_0=2
...

In our output we are getting a bar between two subjects with two spheres at the end.
For example
in the case of subject "1"
we are getting a bar with one end "1_0" and the other end "1_1".

what is this " _0" and "_1" indicating??
How we can identify the two sample category from the plot??
In labels also its showing like this only. How we can customize with emperor?

[Yoshiki Vázquez Baeza]

Dear Prasanth,

To answer your question, to answer this question, you can look at the Options tab, where there's an "edge color selector A" input and and "edge color selector B" input. By default A will be colored white and B will be colored red.

If you look at the Key tab, you'll see that your  samples are suffixed with _0 and with _1, the samples suffixed with _0 are on the white end, and the samples suffixed with _1 are on the red end.

The order of these is as provided in the make_emperor.py command.

Thanks!

Yoshiki.

[Prasanth S]

Hi Yoshiki,

Thanks for the reply.  Our main issue is how to create a mapping file for this input matrices.

we have created a sample_id map like this for transforming the pcoa matrices

Sample_id

1_n =1
2_n=2
...

1_0=1
2_0=2

Do we need to create a sample_id map like this?
we are very new to this platform? we only have this matrix data , no other information's available. Please help
here we are attaching a sample plot .We have to present like this

Thanks
Prasanth

[Yoshiki Vázquez Baeza]

Dear Prasanth,

Procrustes analysis in Emperor, is intended to be used only with matching samples, meaning that you only use one mapping file which describes the samples contained in all the coordinates files.

If the first coordinate file includes samples A, B and C, the second coordinates file should include samples A, B and C as well. Therefore the mapping file only needs to describe the samples A, B and C.

Thanks!

Yoshiki.

[Prasanth S]

HI  Yoshiki Vázquez Baeza,

Thanks again for your reply

As you said

"The order of these is as provided in the make_emperor.py command."

How can we give order,since we can only give the file path?
My folder conataind transformed files for both "n" and "o".
How make_emperor taking these files ? by alphabetic order or they are taking the first one in the folder then second??

Thanks
Prasanth

[Yoshiki Vázquez Baeza]

Hello,

If the filenames in the folder you have are numbered with an underscore and a number, the number will be used to sort the files.

For example coordinates_0.txt, coordinates_1.txt will make it such that the data in _0 will be first and the data in _1 will be the second (with the order as described above).

Thanks!

Yoshiki.

[Prasanth S]

Hi Yoshiki Vázquez Baeza,

Thanks for the reply. It was very helpful.

we have few more doubts regarding this.


1)  How we can change the colour of spheres?
      example:-  In our plot for two transformed pcoa values from samples "0" and "1" ,the subjects "A" ,"B" and "C" are connected as follows.

            A_0----------------A_1
            B_0----------------B_1
            C_0----------------C_1

 we need to colour all "_0" in one colour and "_1" to other colour. How it is possible? Colour by option is not working since we only have sample id values.No other informations available.
 Please help.

[Yoshiki Vázquez Baeza]

Hello,

This is currently not possible in Emperor, since all pairs of samples share metadata, so they are all the same color.

Thanks!

Yoshiki.

[Prasanth S]

Hi Yoshiki,

           Thanks again for your valuable reply.

           We got the graphs. Now as part of the publication purpose,we need to convert this to a png file. But after converting it, the image don't have that much clarity.
           
           We first convert it into pdf and then convert it to png. But we are not getting clarity images.
 
           How you people are presenting this 3D images on your publications??  Please help.

Thanks

Prasanth

[Yoshiki Vázquez Baeza]

Hello,

You can save the plot as an SVG file (see the Options tab) and then convert that SVG file to whatever format you want using Inkscape, for example:

inkscape -f figure_example.svg -A figure_example.pdf

You could also do that using Inkscape's GUI.

Thanks!

Yoshiki.