CEGMA complete/partial depend on number of sequences?

[reema]

Hello Everybody

I have a question related to CEGMA complete and partial score . We have done the assembly by Trinity and then assembled trinity transcripts in more complete transcript set. Now when we find out the number of complete and partial CEGMA score, that makes us a bit confuse. Here's the description:-

Trinity Transcripts [A] [Total 31259] = 202 complete, 229 partial
Assembled Trinity Transcripts [B] [Total 19920]= 186 complete, 213 partial

As the set B has more completed transcript, we are expecting high score for complete and partial as compare to set A [ where the transcripts are fragmented]. We are doubting that this difference in score is because of the total number of transcripts. Set B has lower number of transcripts as compare to set A so set B has lower CEGMA score as compare to Set!. But we are not sure. So my question is :-

1) Is Cegma Complete and partial score also depend on the total number of transcripts/sequences?

Any suggestion would be very helpful.

Many Thanks,
Reema,

[kbradnam]

CEGMA does not care about the number of sequences in your input assembly. CEGMA was designed to work against genome, not transcriptome, assemblies; so you could have just one (very long) sequence in your input and still have 248 complete genes present.

The most parsimonious explanation —*to me anyway*— of your results is that your dataset B contains many incorrectly assembled transcripts compared to dataset A, such that CEGMA can not map as many complete (or partial) genes to the sequences. You could identify the complete genes that are in A and not B, and then BLAST them against B to see if they exist in any form (maybe as strange chimeric transcripts?).