Subject: How can I handle 5 different time series gene expression pattern for the yeast gene SAM3 to infer gene functions for similarities between microarray time series gene expression patterns?
Hi
I am trying to analyze the series GSE3431. Its title is "Logic of the yeast metabolic cycle". Its URL is https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE3431.
I am trying to infer gene functions from similarities and dissimilarities between the trajectories between the expression trajectories of genes.
To trust my results I must understand why I am getting five different expression patterns for the yeast gene with the common name "SAM3".
I tried in vain to find the nucleotide gene sequences for the 5 different SAM3 expression pattern. Their prob-set IDs are
2597_at,
2598_g_at,
2599_s_at,
2600_s_at,
8021_at
The information in each of the 5 lines is identical, except for the last line, which also lists the systematic name of SAM3. How can I chose the correct expression pattern from these 5 possible options? What's the reason that generated these 5 options?
I have attached the pdf file with the gene expression time series trajectories for SAM3. The difference in the magnitude between their expression levels is too big to simply take the average. What is the best way to select the best out of these 5 possible options?
The platform / chip of this dataset is
GPL90 [YG_S98] Affymetrix Yeast Genome S98 Array
Thanks a lot
Thomas Hahn
(Bioinformatics graduate student)
https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE3431
Hi Thomas,
It is often found to map one gene (locus) to multiple probes and one probe to multiple genes (loci). The first case is considered as redundancy and the second one is considered as ambiguity.
According to my knowledge, Those Probes mapping to multiple genes should be discarded. However, one gene mapped to multiple probes is considerable. In this scenario, if you need only raw expression value, take the extreme one (+ve or -ve), given that all the probes of the same gene are showing positive or negative (up / down) expressions. Another practice is to take the average of the expressions of the probes for the same gene. But multiple probes mapping should be discarded if few of them have +ve and few -ve.
Hope this will help.
Please look into the link https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4295278/ (No3CoGP: non-conserved and conserved coexpressed gene pairs)
When multiple probes map to the same gene and if the expression values of these probes vary in a particular condition, we have followed the method used by Dozmorov et al. (2011) to set the expression value of that gene. Let us consider a case that the three probes (P1, P2 and P3) map to the gene G1 and we are considering expression of the gene G1 in two different samples (Sample 1 and Sample 2). Further we consider that the expression of these three probes vary within a sample. If the expression of probe P1 is highest in Sample 1, then the expression of the gene G1 is set to the expression value of the probe P1. On the other hand, if the expression of probe P3 is highest in Sample 2, then we set the expression of gene G1 to the value of expression of probe P3 in Sample 2.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics