Updates - SBAmin.com / Samir Amin

Quarto Demo: R and Python in Action

SBAmin — Sat, 07 Feb 2026 00:00:00 -0500

This post demonstrates mixing R and Python code in a single Quarto document rendered through Jekyll Everything below - tables, plots, and inline computations - is executed at render time.

Claude Code was used to implement quarto integration with Jekyll.¹

R: Summary Table

The mtcars dataset ships with R. Here is a quick summary of the first six rows.

kable(head(mtcars[, 1:6]), caption = "First six rows of mtcars")

Table 1: First six rows of mtcars

	mpg	cyl	disp	hp	drat	wt
Mazda RX4	21.0	6	160	110	3.90	2.620
Mazda RX4 Wag	21.0	6	160	110	3.90	2.875
Datsun 710	22.8	4	108	93	3.85	2.320
Hornet 4 Drive	21.4	6	258	110	3.08	3.215
Hornet Sportabout	18.7	8	360	175	3.15	3.440
Valiant	18.1	6	225	105	2.76	3.460

The dataset contains 32 cars and 11 variables.

R: Scatter Plot

Code

ggplot(mtcars, aes(x = wt, y = mpg, color = factor(cyl))) +
  geom_point(size = 3, alpha = 0.8) +
  geom_smooth(method = "lm", se = FALSE, linewidth = 0.6) +
  scale_color_brewer(palette = "Set1", name = "Cylinders") +
  labs(x = "Weight (1000 lbs)", y = "Miles per Gallon") +
  theme_minimal()

fig-mpg-vs-wt

Figure 1: Fuel economy drops as vehicle weight increases.

See Figure 1 for the relationship between weight and fuel economy.

R: Bar Chart

ggplot(mtcars, aes(x = factor(cyl), fill = factor(cyl))) +
  geom_bar(show.legend = FALSE) +
  scale_fill_brewer(palette = "Pastel1") +
  labs(x = "Cylinders", y = "Count") +
  theme_minimal()

Figure 2: Distribution of cars by cylinder count.

Using column-margins

We know from the first fundamental theorem of calculus that for x in \[*a*, *b*\]:

$$\frac{d}{dx}\left( \int_{a}^{x} f(u)\\du\right)=f(x).$$

Python: Data Wrangling

import pandas as pd
import numpy as np

np.random.seed(42)

df = pd.DataFrame({
    "gene": [f"GENE_{i}" for i in range(1, 11)],
    "log2fc": np.random.normal(0, 2, 10).round(2),
    "pvalue": np.random.uniform(0.0001, 0.1, 10).round(4),
})

df["significant"] = df["pvalue"] < 0.05
print(df.to_string(index=False))

   gene  log2fc  pvalue  significant
 GENE_1    0.99  0.0183         True
 GENE_2   -0.28  0.0184         True
 GENE_3    1.30  0.0305         True
 GENE_4    3.05  0.0525        False
 GENE_5   -0.47  0.0433         True
 GENE_6   -0.47  0.0292         True
 GENE_7    3.16  0.0612        False
 GENE_8    1.53  0.0140         True
 GENE_9   -0.94  0.0293         True
GENE_10    1.09  0.0367         True

Python: Volcano-Style Plot

Using aside class

This is a span that has the class aside which places it in the margin without a footnote number.

import matplotlib.pyplot as plt

np.random.seed(7)
n = 200
log2fc = np.random.normal(0, 1.5, n)
pvals  = 10 ** (-np.abs(log2fc) * np.random.uniform(0.5, 3, n))
sig    = (np.abs(log2fc) > 1) & (pvals < 0.05)

fig, ax = plt.subplots(figsize=(6, 4))
ax.scatter(log2fc[~sig], -np.log10(pvals[~sig]),
           alpha=0.4, s=15, color="grey", label="NS")
ax.scatter(log2fc[sig], -np.log10(pvals[sig]),
           alpha=0.7, s=20, color="crimson", label="Significant")
ax.axhline(-np.log10(0.05), ls="--", lw=0.7, color="black")
ax.axvline(-1, ls="--", lw=0.7, color="black")
ax.axvline( 1, ls="--", lw=0.7, color="black")
ax.set_xlabel("log2 Fold Change")
ax.set_ylabel("-log10(p-value)")
ax.legend(frameon=False)
ax.set_title("Simulated Volcano Plot")
plt.tight_layout()
plt.show()

Figure 3: Simulated differential expression results.

Python: Heatmap

np.random.seed(99)
genes   = [f"Gene{i}" for i in range(1, 9)]
samples = [f"S{j}" for j in range(1, 6)]
mat     = np.random.randn(8, 5)

fig, ax = plt.subplots(figsize=(5, 4))
im = ax.imshow(mat, cmap="RdBu_r", aspect="auto")
ax.set_xticks(range(len(samples)), samples)
ax.set_yticks(range(len(genes)), genes)
fig.colorbar(im, ax=ax, shrink=0.8, label="z-score")

<matplotlib.colorbar.Colorbar object at 0x16de2a810>

ax.set_title("Expression Heatmap")
plt.tight_layout()
plt.show()

Figure 4: Random expression matrix for 8 genes across 5 samples.

Inline Math

The fold change is defined as FC = 2^log₂FC, and significance is assessed at α = 0.05.²

Test a display mathjax

|ψ₁⟩ = a|0⟩ + b|1⟩

Is it O.K.?

Test a display math with equation number

Is it O.K.?

More equations with numbering

\[\begin{align} |\psi_1\rangle &= a|0\rangle + b|1\rangle \\ |\psi_2\rangle &= c|0\rangle + d|1\rangle \end{align}\] \[\begin{align} |\psi_1\rangle &= a|0\rangle + b|1\rangle \\ |\psi_2\rangle &= c|0\rangle + d|1\rangle \end{align}\]

Is it O.K.?

Disable equation numbering

\[\begin{align*} |\psi_1\rangle &= a|0\rangle + b|1\rangle \\ |\psi_2\rangle &= c|0\rangle + d|1\rangle \end{align*}\]

Reenable equation numbering

\[\begin{align} |\psi_1\rangle &= a|0\rangle + b|1\rangle \\ |\psi_2\rangle &= c|0\rangle + d|1\rangle \end{align}\]

Is it O.K.?

Example math with

$\KaTeX$

Let’s test some inline math x , y , x₁ , y₁ .

Now a inline math with special character: ∥ψ⟩ , x^′ , and ∥ψ₁⟩ = a∥0⟩ + b∥1⟩

Example 1:

f(x) = ∫_−∞^∞f̂(ξ) e^{2πiξx*</sup> *dξ}

Example 2:

\[\frac{1}{\Bigl(\sqrt{\phi \sqrt{5}}-\phi\Bigr) e^{\frac25 \pi}} = 1+\frac{e^{-2\pi}} {1+\frac{e^{-4\pi}} {1+\frac{e^{-6\pi}} {1+\frac{e^{-8\pi}} {1+\cdots} } } }\]

Example 3:

f(x) = ∫_−∞^∞f̂(ξ) e^{2πiξx*</sup> *dξ}

Example 4:

\[1 + \frac{q^2}{(1-q)}+\frac{q^6}{(1-q)(1-q^2)}\]

\[\KaTeX\]

See if this E = m**c² works!

Equation 1:

Equation 2:

Source for equations (websites may have been outdated):

Quick R One-Liner

set.seed(123)
sample(letters, 10)

 [1] "o" "s" "n" "c" "j" "r" "k" "e" "x" "y"

Done!

Unlike an earlier post on R + Jekyll integration, the current Quarto-based approach is more flexible, allowing for seamless execution of both R and Python code blocks at render time. ↩
Using katex for math rendering. ↩

tweet author

On being objective and kind

SBAmin — Thu, 06 Mar 2025 00:00:00 -0500

Carl Sagan, in 1990s, unveiled an iconic, Pale Blue Dot picture of the earth. Have a look and see if you can spot what we call as our sweet home! Carl’s a minute and half introduction to this picture is worth listening, especially in times like today when we the people are divided and on two or more sides of that spectrum, we feel insecure but also headstrong in our ideology and way of life.

Pale Blue Dot by NASA - A picture of the earth taken by Voyager 1 from about 6 billion km (3.7 billion miles)

I believe that we all, irrespective of our ideology, have an objective moral compass. The challenge is in having the courage to summon that compass and importantly, not dismiss or drift away when the answer conflicts with our self-interest or ambitions. Adam Smith’s the invisible hand theory¹² stated that how participants in the free market, while acting for their own self-interest, can also inadvertently benefit the society. Although this theory may have lost its merit in today’s disordered world order, our own self-interest and ambitions can - and hopefully should - align with the greater good rather than leaving the other side(s) insecure and vulnerable.

As much as I wish the current administration to follow their moral compass, we being on the other side, too might as well fail to look at ourselves objectively³. We should ask if our professional actions and career ambitions, are aligned to a goodwill of the society or whether we prioritize our self-interest at the cost of compromising on long-term collective outcomes, such as improved health and wealth with decreasing inequality⁴. This does not imply that our self-interest cannot lead to positive societal outcome; rather, a complacent attitude to achieve one’s self-interest can result in mediocrity despite our potential to achieve better outcomes. Unfortunately, amid shrinking funding, rising distrust, and an indolent research culture that fails to overhaul how it should evaluate and incentivize high-impact research, even the most principled scientists face an impossible choice, i.e., to persevere in achieving stated specific aims with utmost scientific rigor or succumb to the pressure to keep the lights on.

Carl Sagan unveils the Pale Blue Dot image Source: @planetarysociety

In the grand cosmic scale, our existence on this sweet home 🌍 is both ephemeral and yet precious. Whether driven by self-interest or altruism, we the people face the same choice — and as this quote reminds me, we can always choose to be kind and still succeed.

You CAN be kind, and still succeed. If you have to compromise who you are to succeed, it’s not worth it.

via AW

Adam Smith’s invisible hand by Peter Foster. ↩
Adam Smith’s invisible hand theory explained - Video by Federal Reserve of Bank of St. Louis. ↩
Look at yourself objectively - Raw Nerve by Aaron Swartz. If blog is not accessible, you may download archived pdf of posts by Aaron Swartz. ↩
Gini coefficient - Wikipedia ↩

tweet author

Getting started with HPC

SBAmin — Fri, 31 Dec 2021 00:00:00 -0500

If you are working in genomics or similar data-intensive field, you’ve likely worked or at least heard of High Performance Computing (HPC) environment or otherwise known as a supercomputer¹. My first experience working with HPC was at Dana-Farber in 2008, right after finishing my clinical internship in India and with practically no experience in programming. I was given a fresh user account to login to a mighty HPC. I was all excited only to realize that I had a little idea on what to do after I follow boilerplate instructions on ssh to HPC, e.g., How should I run ? Why can’t I install R packages or why do certain packages throw cryptic errors?

Early days at @dfcidatascience

Fortunately, resources on working with HPC are now much better and available easily. I thought to add one more and ended up documenting my entire HPC setup guide - step by step - for both CPU and GPU based HPCs.

tl;dr: You can checkout this guide at https://code.sbamin.com/hpc. I hope you find it resourceful and above all, it can help you run your data science tools on HPC.

Best wishes!

Although academic HPCs and supercomputers share core design concepts and job scheduler, e.g., slurm, supercomputers are highly specialized HPCs and often compete among the top500 list. ↩

tweet author

Tips from NCI Center for Cancer Training on applying for K and F awards

SBAmin — Wed, 19 May 2021 00:00:00 -0400

Following are minutes from an AACR 2021 session on NCI Opportunities in Funding: NCI-Sponsored Training Awards and Opportunities. For more on writing K99 grant, you may visit k99.sbamin.com where I’ve detailed my recent experience in submitting K99/R00 application.

AACR 2021 Session Link #paywall 💸

PS: I have not proof-read following minutes for accuracy. Please read respective Funding Opportunity Announcement (FOA) and also contact Program Officer (PO) if you have any questions regarding grant application.

Overview of NCI training awards and opportunities

By Oliver Bogler, @NCI_Bogler

Besides extramural funding (grants awarded to inverstigators across US), NIH also has very large intramural research program.
Available trainig awards from NCI, including new award mechanisms: F99/K00 and early K99
Funding rates are quite good ~20% success rate (5-year average) for most mechanisms. View success rates across all NIH grants

Tips for preparing NCI National Research Service Award (NRSA) fellowship applications

By Mark Damico

Types of awards: F30, F31, F32, and F99/K00
Priority to applicants with training in areas having clear cancer focus and/or whose PI has cancer related R01 or R01-like grants (at least three years in duration and at least 150,000 per year in direct cost) at the time of application and at the time of the award.
Use the correct biosketch version.
- Emphasis why you are the best and the brightest in the personal statement section.
- Highlight any issues that could have impacted your research productivity
Interconnected aims are ok but interdependent aims are not ok.
Preliminary data is not required BUT always know that other applicants may end up submitting preliminary data. So, if you have preliminary data, make sure to include it for a competitive edge.
Emphasis on how you are going to get research training, how does it align to your current and previous training, how will you meet your training goals and objectives, including goals for professional career development in the proposed timeline.
3-5 letters of recommendation and writers must not be directly involved (as collaborators or advisors) in the application.
Don’t despair if you do not get grant into the first submission. Resubmission success rates are always higher than the first submission.
- Make sure to read and re-read summary statement and identify major and minor critiques and how do you plan to address these.
- Then, make sure to talk to your Program Officer (PO) to talk about your summary statement and received their feedback. This is highly encouraged by POs, so don’t be shy but just make sure to carefully read and understand summary statement you would have received.

NCI individual fellowships and career development awards

By Mariam Eljanne

New F99/K00
- Provides continued funding from your graduate research (F99) to post-doctoral (K00) research.
- International students on F-1 visa are eligible to apply
- Must be nominated by a PhD-granting institute: One nominee per institute per year.
- It follows somewhat different format than traditional F fellowship package. Read details under respective FOA link and talk to grant coordinator officer at your institute.
- Average success rate is ~37%.
- See additional resources at F99/K00 NCI webpage

Tips for preparing NCI Career Development (K) Award applications

By Michael K. Schmidt

You may view similar webinar from NCI via following links
Perhaps the most important and the most difficult point in candidate’s background and biosketch is to convince reviewers that you have a potential to make important contributions to the field.
- Many reviewers will look into your past productivity - primarily by number of first-author publications.
- On average, successful applicants have 5 first-author publications but do note that this is not the only criteria reviewers will check. Impact of first-author publication(s) may compensate for <5 number of first-author publications.
Don’t loose sight of the big picture: Why you love the science you do and your goal to become an independent scientist.
Build your team wisely - mentors and collaborators.
Plan ahead and don’t rush into writing and submitting.
Reach out to your mentors, colleagues, and POs for any questions you may have.
Remember to publish, publish, and publish.

Q & A

(I have not covered all of Q & A)

Candidate background and Career Development Plan is equally, if not more, important than Research Strategy in driving overall impact score.
NCI may not award K99 and possibly other training awards despite the top priority score if proposed research has scientific overlap with grant(s) (current or pending but awarded) from your PI or yourself.
- Scientific questions that you are asking should not overlap.
- NCI may ask to look into other grants, including foundation grants, to verify no scientific overlap exists with proposed research.
Mechanisms for postdocs with more than 4-5 years of experience: R50, K22.

tweet author

Writing a K99/R00 grant

SBAmin — Mon, 14 Sep 2020 00:00:00 -0400

K99/R00: Pathway to Independence Award is a NIH funded grant mechanism for a postdoctoral fellow aiming to establish an independent research career following structured postdoctoral training. I have recently - June, 2020 cycle - applied K99/R00 grant. I wrote my experience as a guide that I hope prospective applicants would find resourceful. You can view this guide at k99.sbamin.com

I prefer not to summarize my experience as a collection of a few take-home pointers versus remaining text of trivial value, and rather I found constructive feedback and helpful hints across all of sections of grant writing quite valuable.

Hope you will find it helpful as you hit the next milestone¹ of successfully submitting your K99/R00 application.

Quoting an awesome mentor: “Submitting is a feat on its own. Great job regardless of outcome.” ↩

tweet author

Eulogy: Isaiah Josh Fidler

SBAmin — Sun, 10 May 2020 00:00:00 -0400

Isaiah “Josh” Fidler passed away on 8th May at his Houston, TX home. I did not know him personally but came to know about his work during my graduate class in cancer biology. In 1977, Josh Fidler and Margaret Kripke while working on B16 melanoma cells using C57BL/6 mice conclusively demonstrated¹ that primary tumors - by the process of clonal evolution - are inherently heterogeneous and such preexisting variation enable certain clone(s) or a seed - subpopulation of cancer cells with specific set of variations - to adapt to specific organ(s) or a soil for a metastatic growth.

Nothing in biology is random, there are only processes that we don’t understand.

Isaiah Fidler

Prof. Fidler always liked telling this quote which rhymed well with what Darwin wrote². He stressed that it is the adaptive or Darwinian evolution at-large that is reflected as metastatic heterogeneity, the formidable challenge that is yet to be conquered for better treatments of patients with cancer. Subsequent work by Filder et al. and others strongly supported Stephen Paget’s seminal Seed and Soil theory³ over James Ewing’s counter argument that metastasis is a result of anatomically configured blood and lymphatic supply between primary and metastatic site(s)⁴⁵⁶⁷.

Josh Fidler, image from mdanderson.org

I am trained in computational cancer biology and I have always found appealing to read foundational work on cancer evolution that happened in pre-genomic era⁸⁹¹⁰¹¹¹². Fidler & Kripke’s work was one of those milestones and I was lucky to hear Prof. Fidler twice in Chalk Talk series at the Department of Cancer Biology, MD Anderson.

Video source: https://youtube.com/PLt04fMF6wFwjhfOY0HQxeePUVm0OPVbON

Fidler IJ, Kripke ML. Metastasis results from preexisting variant cells within a malignant tumor. Science 1977;197:893–5. ↩
Excerpt from Charles Darwin’s On The Origin of Species…. ↩
Paget S. The distribution of secondary growths in cancer of the breast. Lancet 1889;133:571–3. https://doi.org/10.1016/S0140-6736(00)49915-0. ↩
Fidler IJ. Tumor heterogeneity and the biology of cancer invasion and metastasis. Cancer Res 1978;38:2651–60. ↩
Hart IR, Fidler IJ. Role of organ selectivity in the determination of metastatic patterns of B16 melanoma. Cancer Res 1980;40:2281–7. ↩
Fidler IJ. The pathogenesis of cancer metastasis: the ‘seed and soil’ hypothesis revisited. Nat Rev Cancer 2003;3:453–8. ↩
Langley RR, Fidler IJ. The seed and soil hypothesis revisited–the role of tumor-stroma interactions in metastasis to different organs. Int J Cancer 2011;128:2527–35. ↩
Milestones in cancer research Nature Milestones ↩
Greaves M. Nothing in cancer makes sense except…. BMC Biol 2018;16:22. ↩
Neutral Theory in Cancer Cell Population Genetics. Mol Biol Evol 2018. ↩
Hunter KW, Amin R, Deasy S, Ha N-H, Wakefield L. Genetic insights into the morass of metastatic heterogeneity. Nat Rev Cancer 2018;18:211–23. ↩
Charlesworth B, Charlesworth D. Population genetics from 1966 to 2016. Heredity 2017;118:2–9. ↩

tweet author

Search NIH grants from command line

SBAmin — Thu, 30 Apr 2020 00:00:00 -0400

NIH RePORTER has an updated API endpoint. Read schema at https://api.reporter.nih.gov I have not yet updated grantfinder script to comply with an updated API.

NIH RePORTER tool is a great tool to query grant records and project summary statements. I find it useful to get a good idea on research projects with shared interest, who’s who in domain experts, and current trends for funding high-impact research. If you prefer command-line over website interface for quick search, NIH provides an excellent API interface to query RePORTER database. I have made a bash utility, grantfinder to query RePORTER API. 🙌

For example, you can search by either PI name (-p) or text query (-q) that can search project terms, project title and abstracts. Remaining arguments are optional.

grantfinder -q "cancer aneuploidy" -s OR -y "2017,2018,2019" -a "NCI,NHGRI"

Demo

Installation

To intall grantfinder, download bash utility and place in PATH, typically in your ~/bin/ folder.

mkdir -p "${HOME}"/bin && \
cd "${HOME}"/bin && \ 
wget https://gist.githubusercontent.com/sbamin/c74d74c67163c08267da616bbead9f37/raw -O grantfinder && \
chmod 755 grantfinder

Since grantfinder requires querying RePORTER API and parsing JSON output, you will need following tools in PATH.

http: Command-line API client¹.
jq: It’s like sed for JSON².
bat: cat like command but with fancy output³. If you do not want this, replace bat with cat in grantfinder script.

Limitations

For now, API limits records to maximum 50 regardless of offset argument (-i). Checkout API documentation⁴ for more options.

tweet author

Learning from Spontaneous Canine Brain Tumors

SBAmin — Mon, 10 Feb 2020 00:00:00 -0500

Our paper focusing on comparative oncology of cross-species gliomas is out today in Cancer Cell. Dogs get spontaneous¹ tumors similar to humans in an immuno-competent environment. We sought to trace molecular life history of gliomas across two species, i.e., to identify common set of somatic alterations and underlying evolutionary conserved mutational processes. We also queried age-related impact on development of glioma in canines vs. human pediatric and adult patients of glioma to show if canine gliomas resemble more of human pediatric or adult gliomas.

We performed genomic, epigenomic, and transcriptomic characterization of sporadic glioma in dogs (n=83), and compared those to available genomic profiling of patients of human pediatric and adult gliomas. Canine gliomas were marked by significantly lower somatic mutation burden than human adult gliomas but similar to human pediatric gliomas. Despite lower somatic mutation burden, gliomas across canine and human pediatric and adult patients were marked by presence of recurrent, hot-spot, somatic mutations in known glioma drivers like PDGFRA, PIK3CA, and IDH1, suggesting convergent evolution of cross-species gliomas. Overall, gliomas in dogs, despite their prevalence in adult dogs, resembled human pediatric gliomas by mutational burden and DNA methylation.

Comparative oncology of spontaneous cancer in companion dogs can yield not only insights into evolutionary origins of cancer but also provide strong rationale to conduct parallel clinical trials in immuno-competent companion dogs for potential benefits to both, humans and their best friend!

For more, please read our paper here². We also provide companion website, https://canineglioma.verhaaklab.com, that hosts extended methods and code to generate figures.

Samirkumar B. Amin, Kevin J. Anderson, C. Elizabeth Boudreau, Emmanuel Martinez-Ledesma, Emre Kocakavuk, Kevin C. Johnson, Floris P. Barthel, Frederick S. Varn, Cynthia Kassab, Xiaoyang Ling, Hoon Kim, Mary Barter, Ching C. Lau, Chew Yee Ngan, Margaret Chapman, Jennifer W. Koehler, James P. Long, Andrew D. Miller, C. Ryan Miller, Brian F. Porter, Daniel R. Rissi, Christina Mazcko, Amy K. LeBlanc, Peter J. Dickinson, Rebecca A. Packer, Amanda R. Taylor, John H. Rossmeisl Jr, Kevin D. Woolard, Amy B. Heimberger, Jonathan M. Levine, Roel G. W. Verhaak.

Comparative Molecular Life History of Spontaneous Canine and Human Gliomas.

Cancer Cell 2020;37:243–57.e7. doi: 10.1016/j.ccell.2020.01.004.

Our work on glioma in dogs and humans is out: https://t.co/jzHhnNnt94. Incredibly proud of @SBAmin for basically conducting a TCGA study by himself.
— Roel Verhaak (@roelverhaak) February 10, 2020

Cross-posted at verhaaklab.com

These are spontaneously occurring and not genetically induced brain tumors in dogs. Similar to human patients with glioma, we receive canine tissue samples for genome sequencing from several veterinary hospitals nationwide. These samples were collected either at the time of surgery for treating canine patients with brain tumor or following medical euthanasia for an inoperable, advanced stage tumor. Visit companion manuscript website for details on sample acquisition. ↩
Full text link is provided by the publisher and valid until March 31, 2020. ↩

tweet author

Michael Nachman on the genomic basis of environmental adaptation in house mice

SBAmin — Thu, 05 Sep 2019 10:30:00 -0400

Speaker: Michael Nachman, PhD¹

LCC Auditorium, The Jackson Laboratory, Bar Harbor, ME

Genetic basis of mice adaptation
- Mice cohort
- Results
Phenotypic Plasticity

Most of traits are polygenic and complex adaptive traits.

Bergmann’s (body size) and Allen’s (extremities) rule in context of adapative traits in cold vs hot climate.
Polygenic traits in humans².
Fisher’s geometric model of adaptation³ - Large mutations with large effect size tends to be disfavored (purifying selection) except in cases with artificial or strong selection, e.g., IGF1 mutation in domestic dogs with taller heights or when population is far from optimum.

Genetic basis of mice adaptation

Study environmental impact on genetic and phenotypic adaptation in wild mice.

Mice cohort

Wild mice cohort: 32 populations, 10 mice per populations.
- inland and not coastal mice
- multiple tissues
- low pass (2-3X) WGS and exome seq (Nimblegen, 15X)
Live mice for developing inbred lines
- 25 lines total, all unrelated to capture genetic and phenotypic variation
- between 15-20th generation
- Not sequenced so far

Results⁴

From wild mice cohort
NY vs FL mice
Genomic analysis
- ~280,000 SNPs from ~19K genes with MAF >5%
- Latent Factor Mixed Model (LFMM) to study GWAS.
- Evidence of population structure but no evidence for isolation by distance, i.e., population bottlenecks were rare.
- ~50% of SNPs follow patterns consistent with adaptation to both hot and cold climate.
- Most outlier SNPs are in potential gene regulatory regions.
- A novel method to identify genes of interest related to adaptation in body mass: cis-eQTL + selection scan + clines in gene expression + mutant phenotype⁵
  - Six genes linked to adaptation.
  - Adam17, Bcat2

Phenotypic Plasticity

Complex traits depends on both, genes and environment.
Does phenotypic plasticity constrain or facilitate adaptive evolution?⁶⁷
- Open end question
Using wild mice cohort: NY vs Brazil mice
Tails grow shorter in cold but when Brazil mice are bred in warm env., they grow tails.
House mice from AZ drink less water than rest of cohort.
- Phenotypic plasticity to water stress based on serum sodium level and weight loss
- Gene expression in the kidney b/w AZ and Canada mice under hydrated and dehydrated mice
  - Most plastic changes in expression are in the same direction as evolved changes in expression.
Work in progress

References

UC Berkeley, Academic/Professional Profile ↩
Boyle, Li, and Pritchard 2017 ↩
Orr 1998 ↩
Phifer-Rixey 2018 ↩
Mack 2018 ↩
Ghalambor 2015 ↩
Corl 2018 ↩

Read more...

tweet author

SNO2018: John Sampson on challenges for immunotherapy against brain tumors

SBAmin — Sat, 17 Nov 2018 09:35:00 -0500

Speaker: John H Sampson, MD, PhD, MBA¹

Biology
Diagnostics
Therapy
Clinical Trials
Solutions

The human mind treats a new idea the same way the body treats a strange protein, it rejects it.

Peter Medawar

Surgery has not done so well with brain tumors².
Same goes for chemotherapies
Tumor treating fields (TTF): Be very careful with trials and long-term outcomes before being adopted in standard of care.

Biology

Immune privilege and sequestration in the CNS needs to be redefined.
Communication between the CNS and immune system exists
Activated T-cells do get through BBB, and can return to systemic circulation.
CNS antigens drain through CSF and circulated to cervical LNs.
Functional lymphatic veins are observed in the CNS.

Diagnostics

Intra-tumoral heterogeneity in primary GBM based on several recent single-cell RNAseq papers³.
Unlike melanoma, targeting with checkpoint inhibitors may not work in GBMs as GBM is more of developmental disease, and blanket immune response may end up killing healthy cells.

Therapy

Factors influencing therapy

Target specificity

Bigner et al. J Neurosurg. 1981

Therapy delivery

EGFR monoclonal antibody can go into CNS - phase 1 trial: Scott et al. PNAS. Important to have specificity as only location where mutant EGFR can be found is in GBM cells and not elsewhere.
Hitchhiking and Trojan horse technique to ship immunoRx drugs in brain.

Heterogeneity

subway map similar to heterogeneity: multiple ways to reach a common target
single-agent therapy will not work in background of heterogeneity.
Efficacy of CART (O’Roukre et al. Sci. Transl Med.)
NeoAgs from K27M mutations for glioma T cell Rx.
Low mutational burden, esp. in pediatric tumors and challenge in using neoAg based approaches.
- Alternate is to use neo-epitope based vaccines (CD8 + CD4 T-cell epitopes) to elicit endogeneous immune response using both CD8 and CD4 mediated immune response.

Clinical Trials

Profit and pricing is not an issue.
Bigger issue is patient accrual and design of trials.
Less stringent and more streamlined approval process of trials from review boards and regulatory authorities.

Solutions

Change to meet new world needs
- Integrate big data analytics in decision making
- data is quite underutilized
Yield signs have changed from yellow (until 1971) to red
- So, that drivers can be more attentive and follow rules.
- Same goes here: No perfect solution but hopefully we adopt to the new world, learn and make decisions based on available information. It’s evolution in its own way.

References

Academic/Professional Profile ↩
Dandy et al. JAMA 1928 ↩
Patel et al. Bernstein lab. ↩

Read more...

tweet author

Updates - SBAmin.com / Samir Amin

Quarto Demo: R and Python in Action

R: Summary Table

R: Scatter Plot

R: Bar Chart

Python: Data Wrangling

Python: Volcano-Style Plot

Python: Heatmap

Inline Math

Test a display mathjax

Test a display math with equation number

More equations with numbering

Disable equation numbering

Reenable equation numbering

Example math with

Source for equations (websites may have been outdated):

Quick R One-Liner

On being objective and kind

Getting started with HPC

Tips from NCI Center for Cancer Training on applying for K and F awards

Overview of NCI training awards and opportunities

Tips for preparing NCI National Research Service Award (NRSA) fellowship applications

NCI individual fellowships and career development awards

Tips for preparing NCI Career Development (K) Award applications

Q & A

Writing a K99/R00 grant

Eulogy: Isaiah Josh Fidler

Search NIH grants from command line

Demo

Installation

Limitations

Learning from Spontaneous Canine Brain Tumors

Michael Nachman on the genomic basis of environmental adaptation in house mice

Genetic basis of mice adaptation

Mice cohort

Results4

Phenotypic Plasticity

References

SNO2018: John Sampson on challenges for immunotherapy against brain tumors

Biology

Diagnostics

Therapy

Target specificity

Therapy delivery

Heterogeneity

Clinical Trials

Solutions

References

Results⁴