Incorporated changes from my proofread of the study section #493

cgreene · 2017-05-21T16:15:20Z

No description provided.

agitter · 2017-05-22T03:42:13Z

I'll review this Monday morning

agapow

LGTM, some typos and suggestions

agapow · 2017-05-22T07:40:07Z

sections/04_study.md

@@ -56,7 +56,7 @@ approaches applied to gene expression data are powerful methods for
 identifying gene signatures that may otherwise be overlooked.
 An additional benefit of unsupervised approaches is that
 ground truth labels, which are often difficult to acquire or are incorrect, are
-nonessential. However, careful interpretation must be performed regarding how
+nonessential. However, careful interpretation must be performed when


"careful interpretation must be performed" sounds way awkward to me. "interpretation must be careful when"?

"the genes that have been aggregated into features must be interpreted carefully"?

reworded 👍

agapow · 2017-05-22T07:42:22Z

sections/04_study.md

-its links to complex disease, which will lead to novel diagnostics and
-therapeutics.
+therapies to correct splicing defects. However, to achieve this we expect that
+methods to interpret the "black box" of deep neural networks and integrate


"integrate this with"?

I like it as is. The "integrate" refers to multiple data sources.

agapow · 2017-05-22T07:43:33Z

sections/04_study.md

-would be very time consuming in a lab setting but was easy to simulate using
-their model. As we learn to better visualize and analyze the hidden nodes within
+base pairs in a sequence and see how the model changed its prediction. Though
+time consuming to assay in a lab, this was easy to simulate the computational


"this was easy to simulate the computational": word missing?

agapow · 2017-05-22T07:44:02Z

sections/04_study.md

-million base pairs upstream or downstream from the affected promoter, on either
-strand, even within the introns of other genes [@doi:10.1038/nrg3458]. They do
+million base pairs upstream or downstream from the affected promoter on either
+strand even within the introns of other genes [@doi:10.1038/nrg3458]. They do


"and even" / "or even"

agapow · 2017-05-22T07:45:56Z

sections/04_study.md

 insights.

 ### Single-cell data

-Single-cell methods are generating extreme excitement as biologists recognize
+Single-cell methods are generating excitement as biologists recognize
 the vast heterogeneity within unicellular species and between cells of the same
 tissue type in the same organism [@tag:Gawad2016_singlecell]. For instance,
 tumor cells and neurons can both harbor extensive somatic variation
 [@tag:Lodato2015_neurons]. Understanding single-cell diversity in all its
 dimensions — genetic, epigenetic, transcriptomic, proteomic, morphologic, and


long dash or double dash? or will either work?

went with double dash. I think that's what we've done elsewhere 👍

agapow · 2017-05-22T07:46:15Z

sections/04_study.md

 specific individual, but also to specific pathological subsets of cells.
 Single-cell methods also promise to uncover a wealth of new biological
 knowledge. A sufficiently large population of single cells will have enough
 representative "snapshots" to recreate timelines of dynamic biological processes.
 If tracking processes over time is not the limiting factor, single-cell
 techniques can provide maximal resolution compared to averaging across all cells
 in bulk tissue, enabling the study of transcriptional bursting with single-cell
-FISH or the heterogeneity of epigenetic patterns with single-cell Hi-C or
+fluorescence in situ hybridization or the heterogeneity of epigenetic patterns with single-cell Hi-C or


italicise in situ?

Agreed, "in situ"

agapow · 2017-05-22T07:46:40Z

sections/04_study.md

@@ -586,23 +576,23 @@ for dealing with batch effects [@tag:Shaham2016_batch_effects].

 Examining populations of single cells can reveal biologically meaningful subsets
 of cells as well as their underlying gene regulatory networks
-[@tag:Gaublomme2015_th17]. Unfortunately, machine learning generally struggles
+[@tag:Gaublomme2015_th17]. Unfortunately, machine learning methods generally struggle
 with imbalanced data — when there are many more examples of class 1 than class 2 —


Looks like a single hyphen not a long dash. Suggest this could all be cleaned up near end with a simple search and replace.

agapow · 2017-05-22T07:49:08Z

sections/04_study.md

 [@tag:Abe]. Then, researchers began to use techniques that could estimate
-relative abundances from an entire sample, which is much faster than classifying
+relative abundances from an entire sample more quickly than classifying


I think "faster" reads better than "more quickly"

agapow · 2017-05-22T07:49:41Z

sections/04_study.md

 [@tag:Word2Vec] in natural language processing) for protein family
 classification have been introduced and classified with a skip-gram neural
 network [@tag:Asgari]. Recurrent neural networks show good performance for
 homology and protein family identification [@tag:Hochreiter @tag:Sonderby].
-Interestingly, Hochreiter, who invented Long Short Term Memory (LSTM), delved
-into homology/protein family classification in 2007, and therefore, deep
-learning is deeply rooted in functional classification methods.

 One of the first techniques of *de novo* genome binning used self-organizing
 maps, a type of neural network [@tag:Abe]. Essinger et al. used Adaptive Resonance Theory


Shift citation to just after Essinger et al.?

agitter

Only minor comments from me and @agapow, then looks good to me.

agitter · 2017-05-22T10:46:07Z

sections/04_study.md

 specific individual, but also to specific pathological subsets of cells.
 Single-cell methods also promise to uncover a wealth of new biological
 knowledge. A sufficiently large population of single cells will have enough
 representative "snapshots" to recreate timelines of dynamic biological processes.
 If tracking processes over time is not the limiting factor, single-cell
 techniques can provide maximal resolution compared to averaging across all cells
 in bulk tissue, enabling the study of transcriptional bursting with single-cell
-FISH or the heterogeneity of epigenetic patterns with single-cell Hi-C or
+fluorescence in situ hybridization or the heterogeneity of epigenetic patterns with single-cell Hi-C or


Agreed, "in situ"

agitter · 2017-05-22T10:48:27Z

sections/04_study.md

+outperforming logistic regression and distance-based outlier detection methods.
+However, they did not benchmark against random forests, which tend to work better
+for imbalanced data, and their data was
+relatively low dimensional. Future work is needed to establish the utility of


In light of #495, I don't see how improvements in image classification tell us anything about cell subset identification. Can we stop the sentence after "cell subset identification."?

agitter · 2017-05-22T10:52:27Z

sections/04_study.md

@@ -56,7 +56,7 @@ approaches applied to gene expression data are powerful methods for
 identifying gene signatures that may otherwise be overlooked.
 An additional benefit of unsupervised approaches is that
 ground truth labels, which are often difficult to acquire or are incorrect, are
-nonessential. However, careful interpretation must be performed regarding how
+nonessential. However, careful interpretation must be performed when


"the genes that have been aggregated into features must be interpreted carefully"?

agitter · 2017-05-22T10:53:06Z

sections/04_study.md

-its links to complex disease, which will lead to novel diagnostics and
-therapeutics.
+therapies to correct splicing defects. However, to achieve this we expect that
+methods to interpret the "black box" of deep neural networks and integrate


I like it as is. The "integrate" refers to multiple data sources.

This build is based on c0cbf63. This commit was created by the following Travis CI build and job: https://travis-ci.org/greenelab/deep-review/builds/234828498 https://travis-ci.org/greenelab/deep-review/jobs/234828499 [ci skip] The full commit message that triggered this build is copied below: Incorporated changes from my proofread of the study section (#493) * initial proofreads up to metagenomics * finish proofread * address comments * address build failure

cgreene added 2 commits May 21, 2017 09:50

initial proofreads up to metagenomics

5e7bd10

finish proofread

4adeb24

cgreene requested a review from agitter May 21, 2017 16:15

agapow approved these changes May 22, 2017

View reviewed changes

agitter approved these changes May 22, 2017

View reviewed changes

cgreene and others added 3 commits May 22, 2017 09:12

address comments

f359052

address build failure

bf6c9e3

Merge branch 'master' into cgreene-study-proofread

4f4eff0

cgreene merged commit c0cbf63 into greenelab:master May 22, 2017

cgreene deleted the cgreene-study-proofread branch May 22, 2017 13:40

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Incorporated changes from my proofread of the study section #493

Incorporated changes from my proofread of the study section #493

cgreene commented May 21, 2017

agitter commented May 22, 2017

agapow left a comment

agapow May 22, 2017

agitter May 22, 2017

cgreene May 22, 2017

agapow May 22, 2017

agitter May 22, 2017

agapow May 22, 2017

cgreene May 22, 2017

agapow May 22, 2017

cgreene May 22, 2017

agapow May 22, 2017

cgreene May 22, 2017

agapow May 22, 2017

agitter May 22, 2017

cgreene May 22, 2017

agapow May 22, 2017

cgreene May 22, 2017

agapow May 22, 2017

cgreene May 22, 2017

agapow May 22, 2017

cgreene May 22, 2017

agitter left a comment

agitter May 22, 2017

agitter May 22, 2017

agitter May 22, 2017

agitter May 22, 2017

Incorporated changes from my proofread of the study section #493

Incorporated changes from my proofread of the study section #493

Conversation

cgreene commented May 21, 2017

agitter commented May 22, 2017

agapow left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

agitter left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment