Number of contigs does not match number in count matrix --- de novo assembly

Question

[nikelle.petrillo]

Hello,

I performed de novo assembly using Trinity. About 493,000 contigs were used to assemble the assembly. After creating a count matrix, we only found 38,000 genes...

Does this mean about 93% of our transcripts are not represented in the count matrix?

Thanks so much for the help, Nikelle

[st.ph.n]

Your 493K transcripts are grouped into clusters by ids. 'gene id: TRINITY_DN1000|c115_g5' encoding 'isoform id: TRINITY_DN1000|c115_g5_i1' any other transcripts in the same cluster will have a different value for '_i#' at the end of the fasta header for that entry. What you have is 493K isoforms, representeing 38k genes.

[GenoMax]

Do you expect your genome to contain 493K genes?

Was an effort made to cluster the sequences with something like CD-HIT? Did you map you reads back to the assembly? Did that mapping look reasonable?

[JC]

are you counting genes or transcripts?