If you really want to study expression profiles, forget Excel. Really.

Instead start learning some software made to handle such data. A very commonly used software for this task is R (www.r-project.org) that can be extended withh 1000s of specialized add-on packages, many to specifically handle microarray data (see www.bioconductor.org). It will take you some time and nerves to learn it, but it is absolutely worth it - not onyl for solving your recent problems but also to improve structured thinking and career options. And R is free, and there are tons and tons of tutorials and books freely available in the internet as well.

However, in Excel you can get the log2 of a value with =LOG(x;2) and the inverse function is the poser to base 2: =2^x or =POWER(2;x)

Heatmaps can not really be made with Excel. You can fiddle around with conditional highlightning of cells, but Excel doesn't let you cluster rows or columns. There are some tools in the internet that will produce heatmaps. But again, better learn R!

Python (https://www.python.org/) & MatPlotLib: http://matplotlib.org/

This will create red-to-green-to-blue color scale:  

a = {maximum value}

b = {minimum value}

c = a - b

x = {actual value}

h = (a - x) / (c * 1.5)

s = 1.0

v = 1.0

rgb = colorsys.hsv_to_rgb(h, s, v)

Example 1: x = a => h = 0 => rgb = (1.0, 0.0, 0.0) = "red".

Example 2: x = b => h = 2/3 => rgb = (0.0, 0.0, 1.0) = "blue".

To make the heatmap, plot the data and set "markerfacecolor" of each data point to the calculated rgb value.

I totally agree with Jochen regarding R.

Just one thing - how can you have an upregulated gene with fold change (FC) of 0.6? If no change is observed, FC will be 1. Usually, people call genes upregulated  with an FC of at least 1.7x and 1/1.7= 0.6x as FC for down-regulated genes.

Log2(FC) is what you get from normalised microarray data since log2 transformation of intensity values is part of the normalisation process. 

Learning to use relevant tools in R is therefore likely to result in more accurate analysis.

With respect to heatmaps, you can use MeV:

http://www.tm4.org/mev.html

There is also a nice guideline on normalising affy microarray .cel files which covers RMAExpress, BioConductor, and AMP:

http://www.tm4.org/normalizing.html

Hope it helps.

I second Jochen and Naser about the use of R, very efficient for statistical analysis of any type of data, including transcriptomic data from micro-array. Although the learning slope of R can be steep at first, you can have a good start at it with graphical interface such as Rstudio : https://www.rstudio.com/.

Nevertheless, if you are new to this kind of analysis, I can warmly recommend MeV. It is VERY user friendly, does not need any prior bioinformatic background, and you can easily prepare your data in excel before importing in MeV. The manual is well made and useful. (http://www.tm4.org/)

As for your specific needs, the transformation of data to Log2 is a one-click action in MeV, and the heatmap is the default presentation of the transcriptomic data.

Good luck for analysing your data.

Dude, I don't wanna repeat what the others already stated. But really, go for R. As Sandrine said, Rstudio is very practical and it's going to be your best friend. Working with big data such as Microarray and RNAseq data using Excel give me goosebumps.

For the purpose of generation of heatmaps R has a myriad of packages. Most of them using ggplot2 (http://ggplot2.org/). It's worth your time spent to learn it.

I hope this helps. Cheers.

Martin has rightly replied your questions.

Use "pheatmap" an R package for generation of heatmap

heatmaps links

http://www.chibi.ubc.ca/matrix2png/

https://discover.nci.nih.gov/cimminer/home.do