Hello,
My collaborator and I had each downloaded and installed MCScanX, so we had one version before and one version after the Dec 2024 update. We are getting important differences in our results between the two versions, even when we use the same settings and input files.
The pre-update version gives this header for the .collinearity file:
############### Parameters ###############
MATCH_SCORE: 50
MATCH_SIZE: 5
GAP_PENALTY: -1
OVERLAP_WINDOW: 5
E_VALUE: 1e-05
MAX GAPS: 25
############### Statistics ###############
Number of collinear genes: 117827, Percentage: 81.22
Number of all genes: 145073
The post-update version gives this:
############### Parameters ###############
MATCH_SCORE: 50
MATCH_SIZE: 5
GAP_PENALTY: -1
OVERLAP_WINDOW: 5
E_VALUE: 1e-05
MAX GAPS: 25
############### Statistics ###############
Number of collinear genes: 117080, Percentage: 80.70
Number of all genes: 145073
Unfortunately, the differences we find between the pre- and post-update results are in the genomic regions we are most interested in-- complex regions that contain multiple, overlapping syntenic calls resulting from a ton of local duplications, inversions, and recombination.
The post-update version calls only the overall synteny across the chromosome, with gaps where that straightforward synteny breaks down in the region. The pre-update software called both this overall, chromosome-level synteny AND multiple small syntenic calls within the region, showing the complex rearrangements.
We have manually checked a handful of the collinearity differences, i.e. the extra calls that are made in the pre-update version. Those we've checked all seem to be correct (i.e. the genes called as small syntenic blocks in the pre-update version are indeed syntenic with their blast hit matches in the other chromosome).
I expect these differences were caused by the 'bug fix' that altered the evalue significance calculation. I understand that the rationale that the file coordinates were being incorrectly used for comparisons in some cases. However, as I've been asking myself whether our pre-update results are real, I don't see any reason why this file coordinate issue would falsely call synteny within our complex region, especially since we manually checked them.
I wanted to flag this as a potential issue, as it significantly (but subtly) changes our results.
Hello,
My collaborator and I had each downloaded and installed MCScanX, so we had one version before and one version after the Dec 2024 update. We are getting important differences in our results between the two versions, even when we use the same settings and input files.
The pre-update version gives this header for the .collinearity file:
############### Parameters ###############
MATCH_SCORE: 50
MATCH_SIZE: 5
GAP_PENALTY: -1
OVERLAP_WINDOW: 5
E_VALUE: 1e-05
MAX GAPS: 25
############### Statistics ###############
Number of collinear genes: 117827, Percentage: 81.22
Number of all genes: 145073
The post-update version gives this:
############### Parameters ###############
MATCH_SCORE: 50
MATCH_SIZE: 5
GAP_PENALTY: -1
OVERLAP_WINDOW: 5
E_VALUE: 1e-05
MAX GAPS: 25
############### Statistics ###############
Number of collinear genes: 117080, Percentage: 80.70
Number of all genes: 145073
Unfortunately, the differences we find between the pre- and post-update results are in the genomic regions we are most interested in-- complex regions that contain multiple, overlapping syntenic calls resulting from a ton of local duplications, inversions, and recombination.
The post-update version calls only the overall synteny across the chromosome, with gaps where that straightforward synteny breaks down in the region. The pre-update software called both this overall, chromosome-level synteny AND multiple small syntenic calls within the region, showing the complex rearrangements.
We have manually checked a handful of the collinearity differences, i.e. the extra calls that are made in the pre-update version. Those we've checked all seem to be correct (i.e. the genes called as small syntenic blocks in the pre-update version are indeed syntenic with their blast hit matches in the other chromosome).
I expect these differences were caused by the 'bug fix' that altered the evalue significance calculation. I understand that the rationale that the file coordinates were being incorrectly used for comparisons in some cases. However, as I've been asking myself whether our pre-update results are real, I don't see any reason why this file coordinate issue would falsely call synteny within our complex region, especially since we manually checked them.
I wanted to flag this as a potential issue, as it significantly (but subtly) changes our results.