need comment on the PSI values

[umranyaman]

Hi @wososa,

Hope you are well. I need your opinion for the alternative splicing events where the T in values are small. Do you think lower T values would cause dPSI getting smaller, but however significant(as their P values are <0.05) to the events where T values are big enough (e.g. >50)? I just realised that dPSI values are not in the range (-5 to +5) for small T values anyway, however P values are <0.05. Do you think this is somehow linked to exon number that gene contains, or the Isoforms diversity of some genes is the reason T values getting bigger, so the dPSI is getting bigger.

So in summary, do you think there should be even lower threshold for dPSI depending on the T values to make it more significant compared to Isoforms with higher T values?

Also I would like to ask the what's the meaning of EX of the alternative splicing event EX.TSS please?

Thanks very much, Best, Umran

[wososa]

Hi @umranyaman ,

I may not understand your question. dPSI is certainly influenced by T values relative to N values. I am doing a PSI-Sigma workshop in early December for about 10 friends (people from splicing field). Would you like to join? If you do, please send me an email to klin@cshl.edu.

The “Ex” means extra splicing events that not included in the gtf file. In other words, the events are novel to the gene annotation.

Best, Woody

[umranyaman]

Thanks very much, I'd definitely join, I sent you an e-mail about it.

So when I apply threshold for "significant" splicing events it is P<0.05 AND |dPSI|>5. For the events with lower T values, I got significant P values but |dPSI|<5 since their T/N smaller due to their T values. For most of the important genes I have low T values in each sample, are the supporting reads(of splice junctions) values dependent on the exon numbers of a gene? Should I reconsider those events as significant or not?

Best, Umran