Forum:Homologous sequences problematic in RNA-seq

Question

[chengye]

As we know, there exist numerous homologous sequences in our Human genomes. This problem has impeded our process in the downstream analysis of RNA-seq, e.g. in the quantification estimations due to the fact that reads derive from these regions must be aligned to many positions. Some algorithms may discard these reads or try to 'rescue' them. However, this problem hasn't been handled satisfactorily. I wonder if some new methods to hurdle this problem has emerged.

Answer 1

[h.mon]

The existence of paralogous gene in the human is not very pervasive, and most of them can be quantified precisely, as the paralogous genes have diverged enough to allow unambiguous mapping and quantification. In addition, software such as RSEM,Salmon or kallisto do satisfactorily handle quantification of multi-mapped reads.

Maybe you are referring to some particular case, like immunoglobulins?

[chengye]

Actually, not only homologous sequences but also widespread repeats in our genome do have similar problems in data analysis due to the alignment with uncertainty.