I'm facing some problems in analyzing the sequence variation of gene paralogs in polyploid species' transcriptome, the denovo assembly using default settings of Trinity software packages produced only one of the two paralogs that should have been annotated. The two paralogs are very similar to each other. Please give me some advice. Thanks in advance.
Hi,
we faced the same problem on durum wheat and developed a specific solution (called homeoSplitter) to improve allo-tetraploid assembly.
In this approach we i) use a standard assembly strategy so that homeolog sequence are potentially assembled in a single config ii) map reads on those contigs iii) use HomeoSplitter to disentangle homeologous sequences based on a maximum likelihood optimization exploiting the differential expression of the two homeologous genes in the RNA that is observed through the reads mapping.
The paper has been published in BMC bioinformatics:
Disentangling homeologous contigs in allo-tetraploid assembly: application to durum wheat.
Vincent Ranwez, Yan Holtz, Gautier Sarah, Morgane Ardisson, Sylvain Santoni,Sylvain Glémin, Muriel Tavaud-Pirra and Jacques David
BMC Bioinformatics 2013, 14(Suppl 15):S15.
http://www.biomedcentral.com/1471-2105/14/S15/S15
The software is freely available on this web page:
http://bioweb.supagro.inra.fr/homeoSplitter/
Do not hesitate to contact me if you face any problem while trying to use this software. Or have any question concerning its relevance for your specific problem.
Note that at the same time an alternative disentangling solution (based on phasing software) has been proposed:
http://www.ncbi.nlm.nih.gov/pubmed/23800085?dopt=Abstract&holding=f1000,f1000m,isrctn
I hope this will help.
It all depends on how similar your paralogs are. When to similar only one will be assembled and chimera will be a problem. A few years back we assembled transcriptomes from an allopolyploid plant using ABySS:
https://www.researchgate.net/publication/221894992_Cutoffs_and_k-mers_implications_from_a_transcriptome_study_in_allopolyploid_plants
The bottom line was that you have to explore the parameter space (k-mer and coverage) to get a more complete picture.
I would love to hear about alternative approaches though.
Article Cutoffs and k-mers: Implications from a transcriptome study ...
Using default parameters in a de novo assembler won't always work well. For this approach, I would suggest greatly increasing the stringency of the assembly so that the two different paralogs will get assembled, instead of merging the two together (which is likely what happened here). I also agree with Oliver that doing a few more assemblies varying k-mer, among other parameters, will likely help. It is surprising how different your assemblies will be by changing these parameters.
Hi,
we faced the same problem on durum wheat and developed a specific solution (called homeoSplitter) to improve allo-tetraploid assembly.
In this approach we i) use a standard assembly strategy so that homeolog sequence are potentially assembled in a single config ii) map reads on those contigs iii) use HomeoSplitter to disentangle homeologous sequences based on a maximum likelihood optimization exploiting the differential expression of the two homeologous genes in the RNA that is observed through the reads mapping.
The paper has been published in BMC bioinformatics:
Disentangling homeologous contigs in allo-tetraploid assembly: application to durum wheat.
Vincent Ranwez, Yan Holtz, Gautier Sarah, Morgane Ardisson, Sylvain Santoni,Sylvain Glémin, Muriel Tavaud-Pirra and Jacques David
BMC Bioinformatics 2013, 14(Suppl 15):S15.
http://www.biomedcentral.com/1471-2105/14/S15/S15
The software is freely available on this web page:
http://bioweb.supagro.inra.fr/homeoSplitter/
Do not hesitate to contact me if you face any problem while trying to use this software. Or have any question concerning its relevance for your specific problem.
Note that at the same time an alternative disentangling solution (based on phasing software) has been proposed:
http://www.ncbi.nlm.nih.gov/pubmed/23800085?dopt=Abstract&holding=f1000,f1000m,isrctn
I hope this will help.
HI, Vincent
You said you assembled the transcriptome first . What kind of software do you use ? Trinity ? Do you think your software will also works on hybrids transcriptome ?
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics