rMATS p value 0 and FDR 0

[Lauren Malsick]

Using rMATS turbo v 4.3.0 I got my list of SE.MATS.JCEC.txt files and I found a gene I wanted to follow up on, however I get p-values of 0 and FDR of 0. I saw on this google group from a few years ago this means I can assume the p-value was < 2.2*10^-16... With this new version of rMATs is this still the same? Or is there a way to figure out the p-value limit? I have actually followed up with PCR data on this gene and this event is happing in my cells so I know it is detecting a significant biological change, just want to make sure I write my results appropriately!

ID
GeneID
geneSymbol
chr
strand
exonStart_0base
exonEnd
upstreamES
upstreamEE
downstreamES
downstreamEE
ID
IJC_SAMPLE_1
SJC_SAMPLE_1
IJC_SAMPLE_2
SJC_SAMPLE_2
IncFormLen
SkipFormLen
PValue
FDR
IncLevel1
IncLevel2
IncLevelDifference
22539
RBM39
NA
chrNW_026521924.1
-
18391394
18391467
18389900
18389950
18391692
18391756
22539
271,232,215
37,33,40
51,57,49
65,46,58
222
149
0
0
0.831,0.825,0.783
0.345,0.454,0.362
0.426

[kutsc...@gmail.com]

Here's the post saying pvalue=0 can be interpreted as less than 2.2e-16:
https://groups.google.com/g/rmats-user-group/c/TW534af62fg/m/tZXBs0Y4BAAJ

The smallest PValue I managed to get from the rMATS statistical model is 1.11022302463e-16

Here is output from: rMATSexe -i input.txt -t 1 -o output.txt -c 0.0001

ID IJC_SAMPLE_1 SJC_SAMPLE_1 IJC_SAMPLE_2 SJC_SAMPLE_2 IncFormLen SkipFormLen PValue

1 100,101,102,103 101,102,103,104 102,103,104,105 271,272,273,274 200 100 1.11022302463e-16
2 100,101,102,103 101,102,103,104 102,103,104,105 272,273,274,275 200 100 0

Both rows have high read counts showing a clear difference between the groups. The first row has a PValue of 1.11e-16. The second row has a few additional read counts and a PValue of 0

The pvalue is computed as (1 - P): https://github.com/Xinglab/rmats-turbo/blob/v4.3.0/rMATS_C/src/myfunc.c#L478

P is a double which is limited in how close it can get to 1 while being less than 1. This post (https://stackoverflow.com/questions/71383519/largest-value-representable-by-a-floating-point-type-smaller-than-1) says that limit is 0.99999999999999988898

1 - 0.99999999999999988898 is very close to the smallest PValue I could get (1.11022302463e-16)

Eric