Chromosome label too close to the chromosome plot.

[GurjotSinghSidhu]

I am trying to plot three homologous chromosomes from related species and I found that chromosome label is too close to the actual chromosome. The instructions mention that I can play with labels, their sizes etc, but I don't know how.

The second chromosome I am plotting is inverted (i.e. it starts from left and goes to right). I had to do this to have better visualization of the synteny. But the label for that chromosome is right on the top of the chromosome itself.

my layout file is here:

# y, xstart, xend, rotation, color, label, va,  bed
 .8,     .2,    .9,       0, #1cbdcc, EleCor_B chr 8B, top, __OUTDIR__/EleCor_B.bed
 .5,     .9,    .2,       0, #934c98, EleInd chr6, top, __OUTDIR__/EleInd.bed
 .2,     .2,    .9,       0, #f7931e, EleCor_A chr 8A, bottom, __OUTDIR__/EleCor_A.bed
# edges
e, 0, 1, __OUTDIR__/EleInd.EleCor_B.anchors.simple
e, 1, 2, __OUTDIR__/EleInd.EleCor_A.anchors.simple

and seqid file is here:

8A
Chr6
8B

I tried this updated layout file:
# y, xstart, xend, rotation, color, label, labelposition, labelfontsize, labelshift, va, bed
 .8,     .2,    .9,       0, #1cbdcc, EleCor_B chr 8B, top, 8, 0.05, __OUTDIR__/EleCor_B.bed
 .5,     .9,    .2,       0, #934c98, EleInd chr6, top, 8, 0.05, __OUTDIR__/EleInd.bed
 .2,     .2,    .9,       0, #f7931e, EleCor_A chr 8A, right, 8, 0.05, bottom, __OUTDIR__/EleCor_A.bed
# edges
e, 0, 1, __OUTDIR__/EleInd.EleCor_B.anchors.simple
e, 1, 2, __OUTDIR__/EleInd.EleCor_A.anchors.simple

But it shows me error that:
FileNotFoundError: [Errno 2] No such file or directory: '8'

which I assume means that its not reading the label position properly. Any help would be appreciated!

[Adamtaranto]

But it shows me error that:
FileNotFoundError: [Errno 2] No such file or directory: '8'

Correct, it is expecting the 8th column to be a bed file path.

Where did you find an example with those extra field headers?

Can you please post the output of jcvi --version and the command that you used to generate the plot above.

[GurjotSinghSidhu]

Thanks for your reply! I am using this JCVI version:
JCVI/1.4.2-foss-2022a

And I used this command to generate my plots:

python -m jcvi.graphics.karyotype seqids layout

Honestly speaking, the example of those extra fields was given by ChatGPT. It even went on to suggest me that there are two kinds of layout formats which JCVI can process: CSV layout (the standard layout) and block layout given below -

# Block-style layout file for full control of labels, fonts, shifts

<chr> EleCor_A_chr8A
xstart=0.185
xend=0.715
ystart=0.8
yend=0.9
rotation=0
color=#f7931e
label="EleCor_A chr 8A"
labelposition=right
labelfontsize=8
labelshift=0.05
va=top
bed=/scratch/your_username/project_folder/EleCor_A.bed

I tried that layout file as well, but it didn't work. But at this point, I believe, it's just making things up.

[tanghaibao]

I don't believe ChatGPT will be able to give a correct layout file 😉

If you have the time, try to go through the tutorial or modify the demo data to your needs:
https://github.com/tanghaibao/jcvi/wiki/MCscan-(Python-version)

[GurjotSinghSidhu]

From the tutorial, I see that we can generate a complex micro-systeny plot where one region from a species is mapping to two regions from two different (or same) species:

I was wondering can this be done for chromosome/scaffold scale macro-synteny plots as well? I am working with two species where one has chromosome level assembly while other has just scaffold level. So, I would like to generate something like this:

Any suggestions on how would a layout file look for that would be appreciated. Thank you!

[Adamtaranto]

The edges are numbered starting from zero in the same order as the contigs in the first section.

To put contigs on the same line give them the same y coord, shift them left and right using the x coord.

It is important that the columns in your blocks file follow the same order as the contigs in your layout file i.e. scaf4, Chr1, scaf1, scaf3, scaf2.

# x,   y, rotation,   ha,     va,   color, ratio,            label
0.5, 0.6,        0, left, center,       m,     1,       scaffold 4
0.5, 0.4,        0, left, center, #fc8d62,     1,  Chr 1
0.3, 0.2,        0, left, center,       m,     1,       scaffold 1
0.5, 0.2,        0, left, center,       m,     1,       scaffold 3
0.7, 0.2,        0, left, center,       m,     1,       scaffold 2
# edges
e, 0, 1
e, 1, 2
e, 1, 3
e, 1, 4

[GurjotSinghSidhu]

Thanks for the detailed response! I am facing a new problem now, hopefully you will be able to help me out! I am plotting a macro-synteny and micro-synteny scale plots for three syntenic chromosomes from three different genomes:

Here is the macro-synteny ribbon plot:

Layout file:

# y, xstart, xend, rotation, color, label, va, bed
.7,    .727, .373,        0, #d12c2c, EraTef_A chr8A, center, __OUTDIR__/EraTef_A.bed
.5,    .705, .395,        0, #00796b, EraTef_B chr8B, center, __OUTDIR__/EraTef_B.bed
.3,      .2,   .9,        0, #f7931e, EleCor_A chr8A, center, __OUTDIR__/EleCor_A.bed
# edges
e, 0, 1, __OUTDIR__/EraTef_A.EraTef_B.anchors.simple
e, 1, 2, __OUTDIR__/EraTef_B.EleCor_A.anchors.simple

Seqids file:

8A
8B
8A

Then I selected a region from the .blocks file (with synteny information for the three genomes) and plotted a micro-synteny plot:

blocks.layout file:

# x,    y, rotation,     ha,     va, color, ratio,            label
0.5,  0.7,      180,   left,    center, #d12c2c,     1,        EraTef_A chr8A
0.5,  0.5,      180,   left,    center, #00796b,     1,        EraTef_B chr8B
0.5,  0.3,      180,   left,    center, #f7931e,     1,          EleCor chr8A
# edges
e, 0, 1
e, 1, 2

So far, it seemed to work perfectly. But then, I also decided the plot the micro-synteny plot at the chromosome scale (without selecting any region from the .blocks file), and I observed that it changes the relative position of labelled genes (in the figure below the red colored arrows indicate the positions where the genes were labelled in the previous figure).

It appears to me that this is due to the fact that I am plotting the chromosomes in reverse order (using 180 rotation). I had to do this as I wanted my micro-synteny plot in the similar layout as my ribbon-plot, but even after using 180 rotation value, gene labels should not be off between different iterations/scales of plotting. Now, I am staring to worry about the analyses I have done before. Hopefully it can be resolved easily. Thanks.

[GurjotSinghSidhu]

@Adamtaranto if you can please look into this issue, I would be grateful. Thanks.

[Adamtaranto]

Please post the exact command you used for the last two micro-synteny plots and provide any input files that were different between the two runs.

[GurjotSinghSidhu]

In summary, the problem is that "During microsynteny plot, if you rotate a chromosome by modifying the 'rotation' parameter in 'blocks.layout' file, the chromosomes will rotate but the gene labels mentioned by '--genelabels' argument in the plotting command will stay at the same place, thus confounding the positional analysis".

Here is one simple example:
Ribbon plot of two chromosomes:

Lets color two regions of interest which will also serve as reference:

Now generating microsynteny plots:
blocks.layout file:

# x,    y, rotation,     ha,     va, color, ratio,            label
0.5,  0.7,     0,   left, center, #00796b,     1,        EraTef_B chr8B
0.5,  0.3,     0,   left, center, #f7931e,     1,          EleCor chr8A
# edges
e, 0, 1

plot generated by:

python -m jcvi.graphics.synteny blocks EraTef_B_EleCor_A.bed blocks.layout --genelabelsize=2 --genelabels=Et_8B_058544.mRNA1,Et_8B_059202.mRNA1,Et_8B_058543.mRNA1,Et_8B_059201.mRNA1,ELECO.r07.8AG0623390.1,ELECO.r07.8AG0623690.1,ELECO.r07.8AG0623710.1,ELECO.r07.8AG0624600.1,ELECO.r07.8AG0625780.1,ELECO.r07.8AG0625790.1

As you can see that the orientation of chromosomes in microsynteny plot is not same as it was in the ribbon plot, so an obvious choice would be to rotate both chromosomes by 180. So, updated blocks.layout file:

# x,    y, rotation,     ha,     va, color, ratio,            label
0.5,  0.7,     180,   left, center, #00796b,     1,        EraTef_B chr8B
0.5,  0.3,     180,   left, center, #f7931e,     1,          EleCor chr8A
# edges
e, 0, 1

As you can see that the orientation of chromosomes here is same as in the ribbon plot thus easing the comparison. But although the orientation of chromosomes got inverted from the previous iteration, the gene labels are still at the same place. This leads to confusion.

Please let me know if you need anything else. Hopefully you will be able to help me out. Thanks.

[GurjotSinghSidhu]

Rotation by +20° and -20°. Note: the gene labels still at the same position as there is no rotation.

[Adamtaranto]

Please update previous post with the commands and layout file used to create the first two images.

[GurjotSinghSidhu]

Sorry for the delay. I got busy although this task was always in back of my mind. Here are all the requested files:

For plotting the entire chromosome:

1. Blocks file (2464 lines):
   three_genomes_eragrostis_chr8.blocks.txt

2. Layout file:
   blocks.layout.txt

3. Bed file (for the three genomes combined):
   EraTef_A.EraTef_B.EleCor_A.bed.txt

4. Command used:

python -m jcvi.graphics.synteny \
               three_genomes_eragrostis_chr8.blocks.txt \
               EraTef_A.EraTef_B.EleCor_A.bed.txt \
               blocks.layout.txt \
               --genelabelsize=2 \
               --genelabels=Et_8A_056826.mRNA1,Et_8A_056829.mRNA1,Et_8A_058067.mRNA1,Et_8A_056828.mRNA1,Et_8B_058544.mRNA1,Et_8B_059202.mRNA1,Et_8B_058543.mRNA1,Et_8B_059201.mRNA1,ELECO.r07.8AG0623390.1,ELECO.r07.8AG0623690.1,ELECO.r07.8AG0623710.1,ELECO.r07.8AG0624600.1,ELECO.r07.8AG0625780.1,ELECO.r07.8AG0625790.1,ELECO.r07.8BG0648950.1,ELECO.r07.8BG0648990.1,ELECO.r07.8BG0651540.1,ELECO.r07.8BG0651610.1,ELECO.r07.8BG0652230.1,ELECO.r07.8BG0652840.1,ELECO.r07.8BG0655070.1,ELECO.r07.8BG0655080.1

5. Plot generated:
   three_genomes_eragrostis_chr8.blocks.pdf

Then I plotted a subset of this chromosome. Basically truncated the .bed file used above and kept the region I was intersted in.

1. Layout file and Bed file: same as above.

2. Blocks file (995 lines):
   three_genomes_eragrostis_chr8_MABGC_region.blocks.txt

3. Command used:

python -m jcvi.graphics.synteny \
               three_genomes_eragrostis_chr8_MABGC_region.blocks.txt \
               EraTef_A.EraTef_B.EleCor_A.bed.txt \
               blocks.layout.txt \
               --genelabelsize=2 \
               --genelabels=Et_8A_056826.mRNA1,Et_8A_056829.mRNA1,Et_8A_058067.mRNA1,Et_8A_056828.mRNA1,Et_8B_058544.mRNA1,Et_8B_059202.mRNA1,Et_8B_058543.mRNA1,Et_8B_059201.mRNA1,ELECO.r07.8AG0623390.1,ELECO.r07.8AG0623690.1,ELECO.r07.8AG0623710.1,ELECO.r07.8AG0624600.1,ELECO.r07.8AG0625780.1,ELECO.r07.8AG0625790.1,ELECO.r07.8BG0648950.1,ELECO.r07.8BG0648990.1,ELECO.r07.8BG0651540.1,ELECO.r07.8BG0651610.1,ELECO.r07.8BG0652230.1,ELECO.r07.8BG0652840.1,ELECO.r07.8BG0655070.1,ELECO.r07.8BG0655080.1

4. Plot generated:
   three_genomes_eragrostis_chr8_MABGC_region.blocks.pdf

And as can be seen in the two plots, the positions of gene labels are not consistent. Hopefully you will be able to find the cause of the issue.

[GurjotSinghSidhu]

@Adamtaranto can you please look into this? if you need more information, please let me know. Thanks!

[Adamtaranto]

I think the label placement is behaving as intended. They are being rotated in-place by the same number of degrees as for the contigs themselves. It would be difficult to have the label maintain a fixed distance and orientation to the end of the contig as the change in position would vary with the length of the contig / how much of the contig is being displayed.

@tanghaibao does that seem right?

I would manually reposition the label in inkscape.

[GurjotSinghSidhu]

Thanks @Adamtaranto for your response. But I don't think that the "gene labels are being rotated in-place by the same number of degrees as for the contigs themselves", indeed, they are not being affected by the angle being used for the contigs. They seem to be static even if you change the chromosome/contig angle or zoom-level, while ideally they should also vary if we vary these parameters.

As per your comment that - "It would be difficult to have the label maintain a fixed distance and orientation to the end of the contig as the change in position would vary with the length of the contig / how much of the contig is being displayed", I think that's the problem we want to solve. The gene labels are already maintaining a fixed distance and orientation irrespective of the ends of contig being displayed. We want to it to vary if the angle or zoom-level parameters of the coting being displayed are varied.

If we have to manually change them in ink space, which version should we be consider the actual? In the example plots I provided, I generated two plots which had difference in gene labels relative to the contig, because of contig zoom-level, so which one of these should we consider true and other fasle?

Thanks a lot! If I can get the labels to show up properly, it will remove a significant bottleneck in my analyses.