Entering edit mode
6 months ago
lorenz
•
0
Hi,
I was running BUSCO using euk and transcriptome mode and getting ~30% missing BUSCOs. I looked at the list of missing BUSCOs, and extracted their sequences from the fasta file inside the lineage directory. blastx of the missing BUSCO sequences using the transcriptome as query found transcripts that have a good match to all of the missing BUSCOs.
My initial understanding is the "missing BUSCOs" are not present in the transcriptome, and therefore the transcriptome is incomplete, or misassembled. But it seems this is not the case? Maybe there is a parameter I missed? Is there another way to assess the "completeness" of a transcriptome?
Thanks!
I think you need to share more info on what you did (e.g. commands, e-values, software versions, logs).
yes that might be a case of different e-values threshold for the blasting. Can you tell us what e-value you considered with your blastx? Also, is your organism distantly related to most sequenced ones? Are you studying some part of the tree of life that are undersampled?
To be fair, there are still "missing" BUSCOs that is still not detected using blastx (~5% missing BUSCO). For the rest, the highest e-value i got from blastx was 9e-04, using uniprot database. Lowest score is 200 and lowest identity is 40%.
Im not really sure how to quantify "undersampled" but there are only 5000 proteins available in ncbi. There is a single genome available but mapping of my transcriptomic reads resulted to just 50% alignment. COI barcoding of our samples shows at most 72% identity (0.28 distance) with the nearest most sequenced organism.