tag:blogger.com,1999:blog-3373377137104117454.post7833957849998341133..comments2020-07-15T05:05:51.482-07:00Comments on Meng's Notes: Simple Enrichment Test -- calculate hypergeometric p-values in RAnonymoushttp://www.blogger.com/profile/00601792748517163575noreply@blogger.comBlogger20125tag:blogger.com,1999:blog-3373377137104117454.post-39715312742476219282019-05-28T21:39:30.307-07:002019-05-28T21:39:30.307-07:00Great post! Thank you (:Great post! Thank you (:Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-22321530439381008652019-03-05T02:38:46.507-08:002019-03-05T02:38:46.507-08:00This comment has been removed by the author.Alfonsohttps://www.blogger.com/profile/09585157410310939224noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-4227107148830262132019-02-05T19:10:35.915-08:002019-02-05T19:10:35.915-08:00This comment has been removed by the author.Danielhttps://www.blogger.com/profile/00933792817497128440noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-19573531339916338702018-07-24T15:43:51.377-07:002018-07-24T15:43:51.377-07:00This has been really helpful to me, thank you!This has been really helpful to me, thank you!Anonymoushttps://www.blogger.com/profile/04659917579003676826noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-51105333151473431532017-10-10T13:00:22.257-07:002017-10-10T13:00:22.257-07:00Yes, it easier to interpret now. Thank you.Yes, it easier to interpret now. Thank you.Kumarnoreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-49741585808773294322017-10-07T19:59:20.456-07:002017-10-07T19:59:20.456-07:00When we set parameter lower.tail=TRUE in phyper (o...When we set parameter lower.tail=TRUE in phyper (or alternative='less' in fisher.test), p-value <= threshold suggests significant depletion. An interpretation of p-value is the probability of observing equal or more depletion when null hypothesis is true. Similar principle applies to enrichment test. Either way null hypothesis remains the same - samples are randomly chosen from population (thus no depletion nor enrichment). Does this make sense?Anonymoushttps://www.blogger.com/profile/00601792748517163575noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-55906163458303522072017-10-05T13:57:51.551-07:002017-10-05T13:57:51.551-07:00Thank you for reply.
My question is: Had P-value b...Thank you for reply.<br />My question is: Had P-value been < threshold in your example, would we call sampling a) depleted and b) enriched, respectively?<br />OR,<br />What is the null hypothesis in both the cases in above example?Kumarnoreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-16953710790658603622017-10-04T20:02:34.818-07:002017-10-04T20:02:34.818-07:00Traditionally we interpret p-value > threshold ...Traditionally we interpret p-value > threshold as INCONCLUSIVE, since null-hypothesis can only be rejected but never be proved. I think the trend is to abandon p-value once for all: http://www.sciencemag.org/news/2017/07/it-will-be-much-harder-call-new-findings-significant-if-team-gets-its-way . http://www.stat.columbia.edu/~gelman/research/unpublished/abandon.pdfAnonymoushttps://www.blogger.com/profile/00601792748517163575noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-66400473093099051042017-10-04T17:46:23.522-07:002017-10-04T17:46:23.522-07:00Useful post.
I have question regarding interpreta...Useful post. <br />I have question regarding interpretation of p-values.<br />In both the cases (i.e. depletion and enrichment), the p-value >0.05 (if 0.05 is my threshold), what would I interpret? I understand that it is neither enriched nor depleted in the above example!! Is that correct? In other words, if p-value would have been less than 0.05, I would say that this sampling is enriched or depleted? ThanksKumarnoreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-3041743040633142272017-09-08T09:05:22.341-07:002017-09-08T09:05:22.341-07:00Great post but the colors make it a little hard to...Great post but the colors make it a little hard to compare fisher to phyper. Slight re-order of the colors would make it more clear.Anonymoushttps://www.blogger.com/profile/08384150192917160697noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-76182667182704028002017-06-18T11:57:00.457-07:002017-06-18T11:57:00.457-07:00Good catch. I fixed it. Thank you!Good catch. I fixed it. Thank you!Anonymoushttps://www.blogger.com/profile/00601792748517163575noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-59485278153226640332017-06-13T12:24:02.991-07:002017-06-13T12:24:02.991-07:00Hi - thanks for posting, but please update
failIn...Hi - thanks for posting, but please update <br />failInPop = 54-hitInPop <br />to<br />failInPop = 52-hitInPop<br />Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-16372915033210088952016-04-13T22:08:37.154-07:002016-04-13T22:08:37.154-07:00Superb summary, Meng, thanks for that.Superb summary, Meng, thanks for that.Adi B.https://www.blogger.com/profile/06512349618799326025noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-6117513393106524632016-01-22T09:40:13.905-08:002016-01-22T09:40:13.905-08:00Your post was actually top-ranked in my search, so...Your post was actually top-ranked in my search, so people are surely reading it! I got into reading your other posts and the one explaining the PCA is fabulous. Thank you very much for putting this much time and effort on explaining these very important--but often misunderstood--concepts on bioinformatics. Please keep posting, I'll make sure to come back!Pablonoreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-82022586780434316232016-01-22T09:39:30.656-08:002016-01-22T09:39:30.656-08:00This comment has been removed by the author.Pablo Ramoshttps://www.blogger.com/profile/15600814748594776089noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-72798490063397159182016-01-22T09:29:01.162-08:002016-01-22T09:29:01.162-08:00Thank you Pablo, for the great suggestions! I woul...Thank you Pablo, for the great suggestions! I would made it better if I know people actually read this. Again, thanks!Anonymoushttps://www.blogger.com/profile/00601792748517163575noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-82460184133034280912016-01-22T08:17:14.651-08:002016-01-22T08:17:14.651-08:00Thanks for the note. Since you begin the post talk...Thanks for the note. Since you begin the post talking about GO term enrichment in a set of genes (reason why I landed here during a google search), I think the example should reflect this analysis instead of the how many diamonds from a set of cards (those of us who don't play cards might not even know how many diamonds should be in a deck to start with ;)). Also, using the same color for different meaning parameters of the two functions can be a bit misleading! All best, PabloPablonoreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-33966935781357642552015-09-21T12:48:06.574-07:002015-09-21T12:48:06.574-07:00Thanks. You're right, flies! This is a slip of...Thanks. You're right, flies! This is a slip of keyboard. Apparently the p-value doesn't match if copy and run the code. I corrected it.Anonymoushttps://www.blogger.com/profile/00601792748517163575noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-39603913719812221472015-09-21T12:06:48.400-07:002015-09-21T12:06:48.400-07:00"We subtract x by 1, when P[X ≥ x] is needed...."We subtract x by 1, when P[X ≥ x] is needed." But you subtract one for both depletion and enrichment tests...flieshttps://www.blogger.com/profile/04438262506920089149noreply@blogger.comtag:blogger.com,1999:blog-3373377137104117454.post-21483593894529576522015-04-16T08:39:34.486-07:002015-04-16T08:39:34.486-07:00Nice note!Nice note!Xianjun Donghttps://www.blogger.com/profile/16940344808359020274noreply@blogger.com