D. williostoni (EPS vs PS)
I performed differential expression analysis by comparing cells in the EPS cluster with cells in PS1, PS2, and PS3. I then took the list of genes from this analysis (EPS-biased, PS-biased, Not Significant) and filtered out the orthologs. The majority of orthologs expressed in the germline were conserved with very few genes moving onto the respective Muller element (Table 1).
Table 1. Number of genes per Muller Element.
index | muller_A | muller_D | muller_E |
---|---|---|---|
conserved | 54 | 114 | 177 |
recent_conserved | 0 | 0 | 0 |
moved_on | 0 | 0 | 0 |
gene_death | 32 | 19 | 30 |
moved_off | 2 | 14 | 6 |
other | 24 | 30 | 26 |
Using these conservation classes I wanted to see if there were differences in PS expression. We can look for enrichment in two ways: (1) we can look at the percent of PS cells expressing genes in each class, (2) we can use contingency tables to test to look for a relationship between expression bias and movement class. If MSCI is driving gene movement, we expect that Muller A (X) and D (Neo-X) will have increased movement off for genes important for spermatogenesis. As a control, Muller E (2) would not show selection.
PS Expression
In Figures 1-3 I look at the proportion of PS cells that expressed genes in the different movement classes. The number of genes in the moved_on and moved_off classes are very low (< 10) which makes interpretation difficult. I decided to group genes into “Selection Favor” and “Selection Against” for all statistical analyses. Here I compare the distribution of the proportion of cells using a Mann-Whitney U test. We see an significant increase in PS expression for genes with “Selection Against” for all three Muller elements.
Mann-Whitney U (Favor vs Against): 0.005195675758492797
Figure 1. Gene movement on/off of Muller element A.
Mann-Whitney U (Favor vs Against): 1.1210518575758348e-17
Figure 2. Gene movement on/off of Muller element D.
Mann-Whitney U (Favor vs Against): 1.3127729254231856e-34
Figure 3. Gene movement on/off of Muller element E.
Enrichment of PS biased genes
Next, we looked at contingency tables to test for a relationship between PS biased expression and selection against. Again, cell counts are very low for some combinations (Table 2,4,6) so I combined counts to make statistical analysis possible (Table 3,5,7). Again there is evidence of a enrichment of genes that underwent “Selection Against” for Muller elements D and E.
Table 2. Number of genes with differential expression on Muller A.
bias | conserved | moved_on | gene_death | moved_off |
---|---|---|---|---|
NS | 4 | 0 | 3 | 1 |
PS | 0 | 0 | 3 | 0 |
EPS | 50 | 0 | 26 | 1 |
Table 3. Number of genes with selection on Muller A.
bias | Selection Favor | Selection Against |
---|---|---|
PS Biased | 0 | 3 |
Not Biased | 54 | 31 |
Fisher’s Exact Test: 0.05453087409783467
Table 4. Number of genes with differential expression on Muller D.
bias | conserved | moved_on | gene_death | moved_off |
---|---|---|---|---|
NS | 9 | 0 | 3 | 4 |
PS | 23 | 0 | 8 | 7 |
EPS | 82 | 0 | 8 | 3 |
Table 5. Number of genes with selection on Muller D.
bias | Selection Favor | Selection Against |
---|---|---|
PS Biased | 23 | 15 |
Not Biased | 91 | 18 |
Fisher’s Exact Test: 0.006088378575605279
Table 6. Number of genes with differential expression on Muller E.
bias | conserved | moved_on | gene_death | moved_off |
---|---|---|---|---|
NS | 19 | 0 | 4 | 0 |
PS | 42 | 0 | 16 | 4 |
EPS | 116 | 0 | 10 | 2 |
Table 6. Number of genes with differential expression on Muller E.
bias | Selection Favor | Selection Against |
---|---|---|
PS Biased | 42 | 20 |
Not Biased | 135 | 16 |
Fisher’s Exact Test: 0.00043644429109212303