[pmiguel]

While pmiguel is right concerning the retrotransposons, I was referring to long transcontigs with either a single ORF covering a minor part (single percentage range) of the contig, or several ORFs at several positions. I've seen such transcontigs in extreme-coverage de novo transcriptome assemblies from plants (with mid- to large-sized genomes), and I don't trust them to be single transcripts, or at least not fully processed functional ones.

This could also be a characteristic of LTR-retrotransposon-derived cDNA. It is probably the case that few of the LTR-retros that compose the majority of mid- large genome-size plants are functionally autonomous. But this would not necessarily prevent them from being transcribed.

LTR-retros that have recently transposed will tend to have very long intron-less ORFs encoding GAG-POL. These retros may actually be functionally autonomous -- able to catalyze their own transposition. But over evolutionary time these ORFs sustain mutations that break them up. Some mutations are no doubt due to the heavy cytosine methylation at CG and CNG sites of repetitive sequence in plants hindering repair of 5-methyl-cytosine deamination events. (Ie, C->U: easy to fix; 5MeC->T: hard to fix. I tend to think of this as plant's "slow motion" form of fungal "RIPing"). There also seem to be lots of deletions that gradually "erode" away the elements as the megayears pass. And, of course, as the LTR retros come to compose a larger portion of the genome, the chances of a new transposition occurring into a previously inserted LTR retrotransposon becomes greater and greater. Hence lots of insertion of elements into other elements creating nested clusters.

Anyway, in certain tissues I think LTR-retrotransposons, even ones that are damaged, probably are expressed. Pollen is probably one such tissue, but there may be others.

--
Phillip

[horvathdp]

So I ran the count of the contigs and identified quite a few that had high numbers of genomic fragments mapping to them. There was no obvious pattern or sequences among the ones with the most hits (although several had hits to more than million fragments each). Also, in consideration of sarvidsson's comments, the median size of these highly over-represented contigs is about 400 bases. I still do however expect that at least a fair subset of my contigs may well represent sequences that were assembled from hnRNA and span introns or have significant amounts of extended 3'UTRs.

So, my next step will be to see if I can identify the nature of these over-represented contigs. Towards that end, I generated a list (exactly 6,666 of them frighteningly enough )of those that had 100X greater coverage than expected (based on my earlier Kmer analysis of the genomic libraries). I am in the process of using BLAT again to generate a set of fragments that map to these transcribed sequences that are highly over-represented in the genome. Once I have these, I will do an assembly and take a closer look to see if I can recognize any obvious transposons or retro elements among them. If pmiguel has any good programs for identifying such (scripts for identifying direct or indirect repeats, or other conserved transposases, etc) and would like to collaborate, I am certainly open to that possibility.

[pmiguel]

If pmiguel has any good programs for identifying such (scripts for identifying direct or indirect repeats, or other conserved transposases, etc) and would like to collaborate, I am certainly open to that possibility.

Actually, I haven't done much more than dabble in this field for many years now. Probably easier for you to contact someone else from Bennetzen lab who would has done this sort of thing more recently. Feel free to email me at [email protected] if you need some contacts.

--
Phillip