I have to identify lncRNAs [A pilot study with 4 samples in total I am using Hisat2 and Stringtie for aligning and to find transcripts respectively.

I have used intially library type mode --fr (unknowingly) and did analysis and after reading some papers and blogs I realised I have to use --rf rather. But when I compared both analysis, I have few inferences but more confusions -

(i) No difference in alignment and alignment statistics after using both library options, IGV gives same view for both, so I cannot distinguish, How to make sense, alignment looks exactly same. ==>How to make inference about stranded library then?

(ii) Only after stringtie merge, I observed with --fr mode, I have more transcript models but less transcript models with --rf. From individual samples, stringtie same number of transcript models using both options. ====> Please explain, how ?

(iii) After using library option in Hisat2, using this option in stringtie will make any difference or not ? I have used library type option in hisat2.

(iv) Apart from above concerns, number of lncRNAs identified were very different when different library options were used for alignment. ===> Not able to conclude with strong evidence/example

Please help to understand these queries ?

Thanks in advance.