Tophat2 error: failed to retrieve right/left read for pair

Question

[boludopublico]

Hi there, In the past I used successfully Tophat2 for RNA-seq mapping in a previous server. Now that Im using a new server, Tophat retrieves the following error at the end of the run:

...
        [FAILED]
Error running /usr/bin/tophat_reports ... [the following is all options/settings/directories] ...
Error: failed to retrieve right read for pair # 1126402 !

Any idea what may be the reason? Im trying to the map same samples as before, so it cannot be the input data. Thanks in advance! diego

[GenoMax]

Do you have a paired-end dataset with mismatched number of reads in two files?

[boludopublico]

Thanks. I do have PE, i didn't check if they have mismatched numbers... it may well be the case. But since I did this analysis in the past with these same files, it didn't occurred to me it could be an issue. I also use --no-discordant --no-mixed options, so I thought 'difficult' reads could be ruled out. I will check!

[GenoMax]

If you do have PE files out of sync you can use repair.sh from BBMap suite to fix the sync. Guide here.

[boludopublico]

Thanks. But not the case, now confirmed. I did try this in PE files pre-filtered to share reads in pairs. :0(

[Chirag Nepal]

Please stop using Tophat https://t.co/Es4ohxOEyx Cole and I developed the method in *2008*. It was greatly improved in TopHat2 then HISAT & HISAT2. There is no reason to use it anymore. I have been saying this for years yet it has more citations this year than last #methodsmatter

— Lior Pachter (@lpachter) December 2, 2017

Why don't you use other better/fast tools like STAR, HiSAT, kallisto and others.

[boludopublico]

Thanks. Im aware of this comment from the developers. Yet, for the current project, I need to match previous results analyzed this way for a publication. I find it difficult to understand how something that worked in the near past will stop doing it now, even if its under "...low maintenance, low support stage...". Unless the dependencies had changed?

[GenoMax]

Have you checked which version you are using on the new server? It most likely would not match the old analysis. Using the latest available version would be advisable since you need to use TopHat.

[WouterDeCoster]

I need to match previous results analyzed this way for a publication

If it's for a publication then you definitely need to use up-to-date software, and if you replicate your results with another tool then it's even better.