[NextGenSeq]

I like CLC Bio (except for it's price). It has tools for virtually any type of molecular biology analysis. The NGS tools are very powerful and work extremely well. The microarray analysis tools are weak but I don't know of another software package which does both.

[luc]

To the OP,

I guess you have to face the fact that CLC, too, has a learning curve.
Yes, you are duplicating your data when importing them into CLC and it would be very nice if one could carry out some more simple data manipulations - like generating a random subset for the data that you have already loaded.
The latest version fortunately has added a batch export tool - which was very much needed.

As mentioned before, the ability to visualize your data immediately at any point of your analysis is the big advantage of tools like CLC and I guess Geneious - even for people comfortable with command line tools.

I have not run into the problems you mentioned and do not understand your question about the assembly.
I did benchmark the CLC aligner with the tools on the www.bioplanet.com/gcat website. With the default settings CLC aligns more reads than both BWA-MEM and Bowtie2, albeit also with slightly higher error rate. Obviously it would be easy to use more stringent aligner settings. BTW, the standard CLC aligner is fast and maps long Moleculo reads just fine (their large gap mapper is rather slow in contrast) .

[GenoMax]

The microarray analysis tools are weak but I don't know of another software package which does both.

GeneSpring and Partek Genomics suite can analyze multiple types of genomic data. They started off as microarray data analysis tools but over the years have added the ability to analyze NGS data. (Small print: In case of GeneSpring the NGS module requires a separate license (there is even a proteomics module), where as Partek has a single license that covers the entire suite).

Just to be clear: GeneSpring and Partek are not end to end solutions (like CLC) (there is a separate program called Partek flow that is like CLC but does need a server to run). One has to do alignments with a suitable aligner independently. GeneSpring and Partek can then ingest the BAM files for further statistical analysis.

[sklages]

Further, would you say that something like IGV sucks because it provides a GUI interface for looking at mapping files? Where do you draw your limits,

rhinoceros is talking about "command line alternatives". IGV has no command line alternative. So for (complex) visualization issues GUIs are more or less necessary. For mapping/assembly stuff you are usually more flexible with command line versions.

[NextGenSeq]

GeneSpring and Partek Genomics suite can analyze multiple types of genomic data. They started off as microarray data analysis tools but over the years have added the ability to analyze NGS data. (Small print: In case of GeneSpring the NGS module requires a separate license (there is even a proteomics module), where as Partek has a single license that covers the entire suite).

Just to be clear: GeneSpring and Partek are not end to end solutions (like CLC) (there is a separate program called Partek flow that is like CLC but does need a server to run). One has to do alignments with a suitable aligner independently. GeneSpring and Partek can then ingest the BAM files for further statistical analysis.

Partek is pretty awful and GeneSpring costs as much as CLC Bio.

[lh3]

My general concern with the commercial software is the lack of 3rd-party evaluation. It is extremely hard not to introduce bias when the developers evaluate their own tools.

[luc]

Since the commercial tools usually have free trials, there should not be any insurmountable problems in evaluating them?
Somehow, I actually assume a similar evaluation bias even with open software tools presented in peer reviewed publications.

My general concern with the commercial software is the lack of 3rd-party evaluation. It is extremely hard not to introduce bias when the developers evaluate their own tools.

[vivek_]

Not really, I think just the number of user test cases and bug reports for open sourced tools is extremely high since they are used much more frequently than commercial suites due to ease of availability. I can't really think of any commercial tools that reach such a wide user base of skilled bioinformaticians.

[jorgezajgla]

clc bio experience and maintenance fees

I have used genomics workbench for a couple of years now. My experience is mixed. The exome analysis tools for cancer genomes were not very useful (and frankly, not user-friendly, since I needed one month with tech support to get a pipeline running). Even the most strict parameters gave 10 -100x additional mutations that those detected by GATK/Mutect. Although I did not make a direct comparison and still validating, most of these extra mutations were probably not real (as visualised with IGV). I have used it for RNASeq with more or less good results, which may be comparable with the tuxedo suite, although I am now in the process of a formal comparison/validation. Nevertheless, we are not buying the maintenance fee, since they have very strange terms and conditions (we need two to three months to get the paperwork for the payment here and, even if they do not provide support or upgrades, they charge you). We are not allowed here to pay retroactively for services not provided (as seems to be the case elsewhere).

[yaximik]

I splurged and eventually purchased CLC GW subscription just to find that it is inferior to Geneious 6, at least in mapping reads to a reference. For both I used default settings when mapping MiSeq reads to hg19, that is low sensitivity/fast. CLC GW completed work faster, but Geneious has an important iteration option, which makes a whole lot of difference. With default 3 iterations, Geneious completed the job just slightly later, yet mapped much more reads. With 10 iterations it took about 2-3x longer, but the difference is drastic. Here is some statistics:

22,515,575 reads, library from modern human genome DNA
CLC mapped 21,829,611; 685,964 not mapped
Geneious mapped 22,135,480 (first iteration); 22,193,629 (10 iterations); 321,946 not mapped

110,840,935 reads made from ancient DNA
CLC mapped 11,125,668 reads; 99,715,267 not mapped
Geneious mapped 12,261,862 (first iteration), 17,528,633 (10 iterations); 93,312,302 not mapped

Geneious 6 is substantially cheaper and works much better (at least for reference mapping) and upgrade to a new version is quite affordable as well.

Both were run on a 64-bit RHEL5.8 machine with two quad-core processors and hyper-threading enabled.

[luc]

I have not tried Geneious. It might have a fantastic read mapping tool.
Nevertheless, single data points, using only default settings of an mapper, are not very indicative of the overall performance.
How many reads are mapped to the correct location? Are the mapping quality scores overall realistic?
Please see for example http://www.bioplanet.com/gcat or Hatem et al. 2013 (http://www.biomedcentral.com/content...105-14-184.pdf) for some more in depth tests.
BTW, did you use CLC 7 ? - it maps considerably faster than previous versions. I have not re-done the GCAT with version 7 though. What I like in CLC, in addition to the visual feedback, is the capability to map really long reads up to 20 kb (works most of the times).

[kmcarr]

I splurged and eventually purchased CLC GW subscription just to find that it is inferior to Geneious 6, at least in mapping reads to a reference. For both I used default settings when mapping MiSeq reads to hg19, that is low sensitivity/fast. CLC GW completed work faster, but Geneious has an important iteration option, which makes a whole lot of difference. With default 3 iterations, Geneious completed the job just slightly later, yet mapped much more reads. With 10 iterations it took about 2-3x longer, but the difference is drastic. Here is some statistics:

22,515,575 reads, library from modern human genome DNA
CLC mapped 21,829,611; 685,964 not mapped
Geneious mapped 22,135,480 (first iteration); 22,193,629 (10 iterations); 321,946 not mapped

110,840,935 reads made from ancient DNA
CLC mapped 11,125,668 reads; 99,715,267 not mapped
Geneious mapped 12,261,862 (first iteration), 17,528,633 (10 iterations); 93,312,302 not mapped

Geneious 6 is substantially cheaper and works much better (at least for reference mapping) and upgrade to a new version is quite affordable as well.

Both were run on a 64-bit RHEL5.8 machine with two quad-core processors and hyper-threading enabled.

Simply counting number of reads mapped is not a very useful metric to compare mapping algorithms. A real comparison requires assesment of the false positive and false negative mapping rates.